info@squid-cache.orgYou can download the FAQ as HTML, PDF, compressed Postscript, plain text, linuxdoc SGML source or as a compressed tar of HTML.
Squid is a high-performance proxy caching server for web clients, supporting FTP, gopher, and HTTP data objects. Unlike traditional caching software, Squid handles all requests in a single, non-blocking, I/O-driven process.
Squid keeps meta data and especially hot objects cached in RAM, caches DNS lookups, supports non-blocking DNS lookups, and implements negative caching of failed requests.
Squid supports SSL, extensive access controls, and full request logging. By using the lightweight Internet Cache Protocol, Squid caches can be arranged in a hierarchy or mesh for additional bandwidth savings.
Squid consists of a main server program squid, a Domain Name System lookup program dnsserver, some optional programs for rewriting requests and performing authentication, and some management and client tools. When squid starts up, it spawns a configurable number of dnsserver processes, each of which can perform a single, blocking Domain Name System (DNS) lookup. This reduces the amount of time the cache waits for DNS lookups.
Squid is derived from the ARPA-funded Harvest project.
Internet object caching is a way to store requested Internet objects (i.e., data available via the HTTP, FTP, and gopher protocols) on a system closer to the requesting site than to the source. Web browsers can then use the local Squid cache as a proxy HTTP server, reducing access time as well as bandwidth consumption.
Harris' Lament says, ``All the good ones are taken."
We needed to distinguish this new version from the Harvest cache software. Squid was the code name for initial development, and it stuck.
Squid is updated often; please see the Squid home page for the most recent versions.
Squid is the result of efforts by numerous individuals from the Internet community. Duane Wessels of the National Laboratory for Applied Network Research (funded by the National Science Foundation) leads code development. Please see the CONTRIBUTORS file for a list of our excellent contributors.
You can download Squid via FTP from the primary FTP site or one of the many worldwide mirror sites.
Many sushi bars also have Squid.
The software is designed to operate on any modern Unix system, and is known to work on at least the following platforms:
For more specific information, please see platforms.html. If you encounter any platform-specific problems, please let us know by registering a entry in our bug database.
Recent versions of Squid will compile and run on Windows/NT with the Cygwin / Mingw packages.
Guido Serassio maintains the native NT port of Squid and is actively working on having the needed changes integrated into the standard Squid distribution. Partially based on earlier NT port by Romeo Anghelache.
LogiSense has ported Squid to Windows NT and sells a supported version. You can also download the source from their FTP site. Thanks to LogiSense for making the code available as required by the GPL terms.
We also have a few other mailing lists which are not strictly Squid-related.
All of our mailing lists have ``-subscribe'' and ``-unsubscribe'' addresses that you must use for subscribe and unsubscribe requests. To unsubscribe from the squid-users list, you send a message to squid-users-unsubscribe@squid-cache.org.
As of version 2.5, Squid can terminate SSL connections. This is perhaps only useful in a surrogate (http accelerator) configuration. You must run configure with --enable-ssl. See https_port in squid.conf for more information.
Squid also supports these encrypted protocols by ``tunelling'' traffic between clients and servers. In this case, Squid can relay the encrypted bits between a client and a server.
Normally, when your browser comes across an https URL, it does one of two things:
The CONNECT method is a way to tunnel any kind of connection through an HTTP proxy. The proxy doesn't understand or interpret the contents. It just passes bytes back and forth between the client and server. For the gory details on tunnelling and the CONNECT method, please see RFC 2817 and Tunneling TCP based protocols through Web proxy servers (expired).
Squid is copyrighted by the University of California San Diego. Squid uses some code developed by others.
Squid is Free Software.
Squid is licensed under the terms of the GNU General Public License.
We think so. Squid uses the Unix time format for all internal time representations. Potential problem areas are in printing and parsing other time representations. We have made the following fixes in to address the year 2000:
Year-2000 fixes were applied to the following Squid versions:
Patches:
Squid-2.2 and earlier versions have a New Year bug. This is not strictly a Year-2000 bug; it would happen on the first day of any year.
Yep. Please see the commercial support page.
The following people have made contributions to this document:
Please send corrections, updates, and comments to: squid-faq@squid-cache.org.
This document is copyrighted (2000) by Duane Wessels.
This document was written in SGML and converted with the SGML-Tools package.
Most current version of this document can always be found at http://www.squid-cache.org/Doc/FAQ/ in HTML, Plain Text, Postscript and SGML formats.
It is easier for us if you send us text which is close to "correct" SGML. The SQUID FAQ currently uses the LINUXDOC DTD. Its probably easiest to follow examples in the this file. Here are the basics:
Use the <url> tag for links, instead of HTML <A HREF ...>
<url url="http://www.squid-cache.org" name="Squid Home Page">
Use <em> for emphasis, config options, and pathnames:
<em>usr/local/squid/etc/squid.conf</em> <em/cache_peer/
Here is how you do lists:
<itemize> <item>foo <item>bar </itemize>
Use <verb>, just like HTML's <PRE> to show unformatted text.
You must download a source archive file of the form squid-x.y.z-src.tar.gz (eg, squid-1.1.6-src.tar.gz) from the Squid home page, or. the Squid FTP site. Context diffs are available for upgrading to new versions. These can be applied with the patch program (available from the GNU FTP site).
For Squid-1.0 and Squid-1.1 versions, you can just type make from the top-level directory after unpacking the source files. For example:
% tar xzf squid-1.1.21-src.tar.gz % cd squid-1.1.21 % make
For Squid-2 you must run the configure script yourself before running make:
% tar xzf squid-2.0.RELEASE-src.tar.gz % cd squid-2.0.RELEASE % ./configure % make
To compile Squid, you will need an ANSI C compiler. Almost all modern Unix systems come with pre-installed compilers which work just fine. The old SunOS compilers do not have support for ANSI C, and the Sun compiler for Solaris is a product which must be purchased separately.
If you are uncertain about your system's C compiler, The GNU C compiler is available at the GNU FTP site. In addition to gcc, you may also want or need to install the binutils package.
You will need Perl installed on your system.
The developers do not have the resources to make pre-compiled binaries available. Instead, we invest effort into making the source code very portable. Some people have made binary packages available. Please see our Platforms Page.
The SGI Freeware site has pre-compiled packages for SGI IRIX.
Squid binaries for FreeBSD on Alpha and Intel.
Squid binaries for NetBSD on everything
Gurkan Sengun has some Sparc/Solaris packages available.
You need the patch program. You should probably duplicate the
entire directory structure before applying the patch. For example, if
you are upgrading from squid-1.1.10 to 1.1.11, you would run
these commands:
cd squid-2.5.STABLE3 mkdir ../squid-2.5.STABLE4 find . -depth -print | cpio -pdv ../squid-1.1.11 cd ../squid-1.1.11 patch -p1 < /tmp/squid-2.5.STABLE3-STABLE4.diff or alternatively cp -rl squid-2.5.STABLE3 squid-2.5.STABLE4 cd squid-2.5.STABLE4 zcat /tmp/squid-2.5.STABLE3-STABLE4.diff.gz | patch -p1After the patch has been applied, you must rebuild Squid from the very beginning, i.e.:
make distclean ./configure ... make make install
If your patch program seems to complain or refuses to work,
you should get a more recent version, from the
GNU FTP site, for example.
The configure script can take numerous options. The most
useful is --prefix to install it in a different directory.
The default installation directory is /usr/local/squid/. To
change the default, you could do:
% cd squid-x.y.z % ./configure --prefix=/some/other/directory/squid
Type
% ./configure --helpto see all available options. You will need to specify some of these options to enable or disable certain features. Some options which are used often include:
--prefix=PREFIX install architecture-independent files in PREFIX
[/usr/local/squid]
--enable-dlmalloc[=LIB] Compile & use the malloc package by Doug Lea
--enable-gnuregex Compile GNUregex
--enable-splaytree Use SPLAY trees to store ACL lists
--enable-xmalloc-debug Do some simple malloc debugging
--enable-xmalloc-debug-trace
Detailed trace of memory allocations
--enable-xmalloc-statistics
Show malloc statistics in status page
--enable-carp Enable CARP support
--enable-async-io Do ASYNC disk I/O using threads
--enable-icmp Enable ICMP pinging
--enable-delay-pools Enable delay pools to limit bandwith usage
--enable-mem-gen-trace Do trace of memory stuff
--enable-useragent-log Enable logging of User-Agent header
--enable-kill-parent-hack
Kill parent on shutdown
--enable-snmp Enable SNMP monitoring
--enable-cachemgr-hostname[=hostname]
Make cachemgr.cgi default to this host
--enable-arp-acl Enable use of ARP ACL lists (ether address)
--enable-htpc Enable HTCP protocol
--enable-forw-via-db Enable Forw/Via database
--enable-cache-digests Use Cache Digests
see http://www.squid-cache.org/Doc/FAQ/FAQ-16.html
--enable-err-language=lang
Select language for Error pages (see errors dir)
by Kevin Sartorelli and Andreas Doering.
Probably you've recently installed bind 8.x. There is a mismatch between the header files and DNS library that Squid has found. There are a couple of things you can try.
First, try adding -lbind to XTRA_LIBS in src/Makefile.
If -lresolv is already there, remove it.
If that doesn't seem to work, edit your arpa/inet.h file and comment out the following:
#define inet_addr __inet_addr #define inet_aton __inet_aton #define inet_lnaof __inet_lnaof #define inet_makeaddr __inet_makeaddr #define inet_neta __inet_neta #define inet_netof __inet_netof #define inet_network __inet_network #define inet_net_ntop __inet_net_ntop #define inet_net_pton __inet_net_pton #define inet_ntoa __inet_ntoa #define inet_pton __inet_pton #define inet_ntop __inet_ntop #define inet_nsap_addr __inet_nsap_addr #define inet_nsap_ntoa __inet_nsap_ntoa
If you have source for BIND, you can modify it as indicated in the diff below. It causes the global variable _dns_ttl_ to be set with the TTL of the most recent lookup. Then, when you compile Squid, the configure script will look for the _dns_ttl_ symbol in libresolv.a. If found, dnsserver will return the TTL value for every lookup.
This hack was contributed by Endre Balint Nagy.
diff -ru bind-4.9.4-orig/res/gethnamaddr.c bind-4.9.4/res/gethnamaddr.c
--- bind-4.9.4-orig/res/gethnamaddr.c Mon Aug 5 02:31:35 1996
+++ bind-4.9.4/res/gethnamaddr.c Tue Aug 27 15:33:11 1996
@@ -133,6 +133,7 @@
} align;
extern int h_errno;
+int _dns_ttl_;
#ifdef DEBUG
static void
@@ -223,6 +224,7 @@
host.h_addr_list = h_addr_ptrs;
haveanswer = 0;
had_error = 0;
+ _dns_ttl_ = -1;
while (ancount-- > 0 && cp < eom && !had_error) {
n = dn_expand(answer->buf, eom, cp, bp, buflen);
if ((n < 0) || !(*name_ok)(bp)) {
@@ -232,8 +234,11 @@
cp += n; /* name */
type = _getshort(cp);
cp += INT16SZ; /* type */
- class = _getshort(cp);
- cp += INT16SZ + INT32SZ; /* class, TTL */
+ class = _getshort(cp);
+ cp += INT16SZ; /* class */
+ if (qtype == T_A && type == T_A)
+ _dns_ttl_ = _getlong(cp);
+ cp += INT32SZ; /* TTL */
n = _getshort(cp);
cp += INT16SZ; /* len */
if (class != C_IN) {
And here is a patch for BIND-8:
*** src/lib/irs/dns_ho.c.orig Tue May 26 21:55:51 1998
--- src/lib/irs/dns_ho.c Tue May 26 21:59:57 1998
***************
*** 87,92 ****
--- 87,93 ----
#endif
extern int h_errno;
+ int _dns_ttl_;
/* Definitions. */
***************
*** 395,400 ****
--- 396,402 ----
pvt->host.h_addr_list = pvt->h_addr_ptrs;
haveanswer = 0;
had_error = 0;
+ _dns_ttl_ = -1;
while (ancount-- > 0 && cp < eom && !had_error) {
n = dn_expand(ansbuf, eom, cp, bp, buflen);
if ((n < 0) || !(*name_ok)(bp)) {
***************
*** 404,411 ****
cp += n; /* name */
type = ns_get16(cp);
cp += INT16SZ; /* type */
! class = ns_get16(cp);
! cp += INT16SZ + INT32SZ; /* class, TTL */
n = ns_get16(cp);
cp += INT16SZ; /* len */
if (class != C_IN) {
--- 406,416 ----
cp += n; /* name */
type = ns_get16(cp);
cp += INT16SZ; /* type */
! class = _getshort(cp);
! cp += INT16SZ; /* class */
! if (qtype == T_A && type == T_A)
! _dns_ttl_ = _getlong(cp);
! cp += INT32SZ; /* TTL */
n = ns_get16(cp);
cp += INT16SZ; /* len */
if (class != C_IN) {
cache_cf.c: In function `parseConfigFile': cache_cf.c:1353: yacc stack overflow before `token' ...
You may need to upgrade your gcc installation to a more recent version. Check your gcc version with
gcc -vIf it is earlier than 2.7.2, you might consider upgrading.
The following error occurs on Solaris systems using gcc when the Solaris C compiler is not installed:
/usr/bin/rm -f libmiscutil.a /usr/bin/false r libmiscutil.a rfc1123.o rfc1738.o util.o ... make[1]: *** [libmiscutil.a] Error 255 make[1]: Leaving directory `/tmp/squid-1.1.11/lib' make: *** [all] Error 1Note on the second line the /usr/bin/false. This is supposed to be a path to the ar program. If configure cannot find ar on your system, then it substitues false.
To fix this you either need to:
Please check the page of platforms on which Squid is known to compile. Your problem might be listed there together with a solution. If it isn't listed there, mail us what you are trying, your Squid version, and the problems you encounter.
Warnings are usually not a big concern, and can be common with software designed to operate on multiple platforms. If you feel like fixing compile-time warnings, please do so and send us the patches.
by Doug Nazar
In order in compile squid, you need to have a reasonable facsimile of a Unix system installed. This includes bash, make, sed, emx, various file utilities and a few more. I've setup a TVFS drive that matches a Unix file system but this probably isn't strictly necessary.
I made a few modifications to the pristine EMX 0.9d install.
You will need to run scripts/convert.configure.to.os2 (in the Squid source distribution) to modify the configure script so that it can search for the various programs.
Next, you need to set a few environment variables (see EMX docs for meaning):
export EMXOPT="-h256 -c" export LDFLAGS="-Zexe -Zbin -s"
Now you are ready to configure squid:
./configure
Compile everything:
make
and finally, install:
make install
This will by default, install into /usr/local/squid. If you wish to install somewhere else, see the --prefix option for configure.
Now, don't forget to set EMXOPT before running squid each time. I recommend using the -Y and -N options.
There are no hard-and-fast rules. The most important resource for Squid is physical memory. Your processor does not need to be ultra-fast. Your disk system will be the major bottleneck, so fast disks are important for high-volume caches. Do not use IDE disks if you can help it.
In late 1998, if you are buying a new machine for a cache, I would recommend the following configuration:
Also, see Squid Sizing for Intel Platforms by Martin Hamilton This is a very nice page summarizing system configurations people are using for large Squid caches.
After compiling Squid, you can install it with this simple command:
% make installIf you have enabled the ICMP features then you will also want to type
% su # make install-pinger
After installing, you will want to edit and customize the squid.conf file. By default, this file is located at /usr/local/squid/etc/squid.conf.
Also, a QUICKSTART guide has been included with the source distribution. Please see the directory where you unpacked the source archive.
The squid.conf file defines the configuration for squid. the configuration includes (but not limited to) HTTP port number, the ICP request port number, incoming and outgoing requests, information about firewall access, and various timeout information.
Yes, after you make install, a sample squid.conf file will
exist in the ``etc" directory under the Squid installation directory.
The sample squid.conf file contains comments explaining each option.
First you need to make your Squid configuration. The Squid configuration can be found in /usr/local/squid/etc/squid.conf and by default includes documentation on all directives.
In the Suqid distribution there is a small QUICKSTART guide indicating which directives you need to look closer at and why. At a absolute minimum you need to change the http_access configuration to allow access from your clients.
To verify your configuration file you can use the -k parse option
% /usr/local/squid/sbin/squid -k parseIf this outputs any errors then these are syntax errors or other fatal misconfigurations and needs to be corrected before you continue. If it is silent and immediately gives back the command promt then your squid.conf is syntactically correct and could be understood by Squid.
After you've finished editing the configuration file, you can start Squid for the first time. The procedure depends a little bit on which version you are using.
First, you must create the swap directories. Do this by running Squid with the -z option:
% /usr/local/squid/sbin/squid -z
NOTE: If you run Squid as root then you may need to first create /usr/local/squid/var/logs and your cache_dir directories and assign ownership of these to the cache_effective_user configured in your squid.conf.
Once the creation of the cache directories completes, you can start Squid and try it out. Probably the best thing to do is run it from your terminal and watch the debugging output. Use this command:
% /usr/local/squid/sbin/squid -NCd1If everything is working okay, you will see the line:
Ready to serve requests.If you want to run squid in the background, as a daemon process, just leave off all options:
% /usr/local/squid/sbin/squid
NOTE: depending on which http_port you select you may need to start squid as root (http_port <1024).
NOTE: In Squid-2.4 and earlier Squid was installed in bin by default, not sbin.
Squid-2 has a restart feature built in. This greatly simplifies starting Squid and means that you don't need to use RunCache or inittab. At the minimum, you only need to enter the pathname to the Squid executable. For example:
/usr/local/squid/sbin/squid
Squid will automatically background itself and then spawn a child process. In your syslog messages file, you should see something like this:
Sep 23 23:55:58 kitty squid[14616]: Squid Parent: child process 14617 startedThat means that process ID 14563 is the parent process which monitors the child process (pid 14617). The child process is the one that does all of the work. The parent process just waits for the child process to exit. If the child process exits unexpectedly, the parent will automatically start another child process. In that case, syslog shows:
Sep 23 23:56:02 kitty squid[14616]: Squid Parent: child process 14617 exited with status 1 Sep 23 23:56:05 kitty squid[14616]: Squid Parent: child process 14619 started
If there is some problem, and Squid can not start, the parent process will give up after a while. Your syslog will show:
Sep 23 23:56:12 kitty squid[14616]: Exiting due to repeated, frequent failuresWhen this happens you should check your syslog messages and cache.log file for error messages.
When you look at a process (ps command) listing, you'll see two squid processes:
24353 ?? Ss 0:00.00 /usr/local/squid/bin/squid 24354 ?? R 0:03.39 (squid) (squid)The first is the parent process, and the child process is the one called ``(squid)''. Note that if you accidentally kill the parent process, the child process will not notice.
If you want to run Squid from your termainal and prevent it from backgrounding and spawning a child process, use the -N command line option.
/usr/local/squid/bin/squid -N
On systems which have an /etc/inittab file (Digital Unix, Solaris, IRIX, HP-UX, Linux), you can add a line like this:
sq:3:respawn:/usr/local/squid/sbin/squid.sh < /dev/null >> /tmp/squid.log 2>&1We recommend using a squid.sh shell script, but you could instead call Squid directly with the -N option and other options you may require. A sameple squid.sh script is shown below:
#!/bin/sh
C=/usr/local/squid
PATH=/usr/bin:$C/bin
TZ=PST8PDT
export PATH TZ
# User to notify on restarts
notify="root"
# Squid command line options
opts=""
cd $C
umask 022
sleep 10
while [ -f /var/run/nosquid ]; do
sleep 1
done
/usr/bin/tail -20 $C/logs/cache.log \
| Mail -s "Squid restart on `hostname` at `date`" $notify
exec bin/squid -N $opts
On BSD-ish systems, you will need to start Squid from the ``rc'' files, usually /etc/rc.local. For example:
if [ -f /usr/local/squid/sbin/squid ]; then
echo -n ' Squid'
/usr/local/squid/sbin/squid
fi
Squid ships with a init.d type startup script in contrib/squid.rc which works on most init.d type systems. Or you can write your own using any normal init.d script found in your system as template and add the start/stop fragments shown below.
Start:
/usr/local/squid/sbin/squid
Stop:
/usr/local/squid/sbin/squid -k shutdown
n=120
while /usr/local/squid/sbin/squid -k check && [ $n -gt 120 ]; do
sleep 1
echo -n .
n=`expr $n - 1`
done
You can use the squidclient program:
% squidclient http://www.netscape.com/ > test
There are other command-line HTTP client programs available as well. Two that you may find useful are wget and echoping.
Another way is to use Squid itself to see if it can signal a running Squid process:
% squid -k checkAnd then check the shell's exit status variable.
Also, check the log files, most importantly the access.log and cache.log files.
These are the command line options for Squid-2:
Specify an alternate port number for incoming HTTP requests. Useful for testing a configuration file on a non-standard port.
Debugging level for ``stderr'' messages. If you use this option, then debugging messages up to the specified level will also be written to stderr.
Specify an alternate squid.conf file instead of the pathname compiled into the executable.
Prints the usage and help message.
Sends a HUP signal, which causes Squid to re-read its configuration files.
Sends an USR1 signal, which causes Squid to rotate its log files. Note, if logfile_rotate is set to zero, Squid still closes and re-opens all log files.
Sends a TERM signal, which causes Squid to wait briefly for current connections to finish and then exit. The amount of time to wait is specified with shutdown_lifetime.
Sends an INT signal, which causes Squid to shutdown immediately, without waiting for current connections.
Sends a KILL signal, which causes the Squid process to exit immediately, without closing any connections or log files. Use this only as a last resort.
Sends an USR2 signal, which causes Squid to generate full debugging messages until the next USR2 signal is recieved. Obviously very useful for debugging problems.
Sends a ``ZERO'' signal to the Squid process. This simply checks whether or not the process is actually running.
Send debugging (level 0 only) message to syslog.
Specify an alternate port number for ICP messages. Useful for testing a configuration file on a non-standard port.
Prints the Squid version.
Creates disk swap directories. You must use this option when installing Squid for the first time, or when you add or modify the cache_dir configuration.
Do not make initial DNS tests. Normally, Squid looks up some well-known DNS hostnames to ensure that your DNS name resolution service is working properly.
If the swap.state logs are clean, then the cache is rebuilt in the ``foreground'' before any requests are served. This will decrease the time required to rebuild the cache, but HTTP requests will not be satisified during this time.
Do not automatically become a background daemon process.
Do not set the SO_REUSEADDR option on sockets.
Enable virtual host support for the httpd-accelerator mode. This is identical to writing httpd_accel_host virtual in the config file.
Enable full debugging while parsing the config file.
Return ICP_OP_MISS_NOFETCH instead of ICP_OP_MISS while the swap.state file is being read. If your cache has mostly child caches which use ICP, this will allow your cache to rebuild faster.
Squid is a single process application and can not make use of SMP. If you want to make Squid benefit from a SMP system you will need to run multiple instances of Squid and find a way to distribute your users on the different Squid instances just as if you had multiple Squid boxes.
Having two CPUs is indeed nice for running other CPU intensive tasks on the same server as the proxy, such as if you have a lot of logs and need to run various statistics collections during peak hours.
The authentication and group helpers barely use any CPU and does not benefit from dual-CPU configuration.
RAID1 is fine, and so are separate drives.
RAID0 (striping) with Squid only gives you the drawback that if you lose one of the drives the whole stripe set is lost. There is no benefit in performance as Squid already distributes the load on the drives quite nicely.
Squid is the worst case application for RAID5, whether hardware or software, and will absolutely kill the performance of a RAID5. Once the cache has been filled Squid uses a lot of small random writes which the worst case workload for RAID5, effectively reducing write speed to only little more than that of one single drive.
Generally seek time is what you want to optimize for Squid, or more precisely the total amount of seeks/s your system can sustain. Choosing the right RAID solution generally decreases the amount of seeks/s your system can sustain significantly.
To place your cache in a hierarchy, use the cache_peer
directive in squid.conf to specify the parent and sibling
nodes.
For example, the following squid.conf file on
childcache.example.com configures its cache to retrieve
data from one parent cache and two sibling caches:
# squid.conf - On the host: childcache.example.com
#
# Format is: hostname type http_port udp_port
#
cache_peer parentcache.example.com parent 3128 3130
cache_peer childcache2.example.com sibling 3128 3130
cache_peer childcache3.example.com sibling 3128 3130
The cache_peer_domain directive allows you to specify that
certain caches siblings or parents for certain domains:
# squid.conf - On the host: sv.cache.nlanr.net
#
# Format is: hostname type http_port udp_port
#
cache_peer electraglide.geog.unsw.edu.au parent 3128 3130
cache_peer cache1.nzgate.net.nz parent 3128 3130
cache_peer pb.cache.nlanr.net parent 3128 3130
cache_peer it.cache.nlanr.net parent 3128 3130
cache_peer sd.cache.nlanr.net parent 3128 3130
cache_peer uc.cache.nlanr.net sibling 3128 3130
cache_peer bo.cache.nlanr.net sibling 3128 3130
cache_peer_domain electraglide.geog.unsw.edu.au .au
cache_peer_domain cache1.nzgate.net.nz .au .aq .fj .nz
cache_peer_domain pb.cache.nlanr.net .uk .de .fr .no .se .it
cache_peer_domain it.cache.nlanr.net .uk .de .fr .no .se .it
cache_peer_domain sd.cache.nlanr.net .mx .za .mu .zm
The configuration above indicates that the cache will use
pb.cache.nlanr.net and it.cache.nlanr.net
for domains uk, de, fr, no, se and it, sd.cache.nlanr.net
for domains mx, za, mu and zm, and cache1.nzgate.net.nz
for domains au, aq, fj, and nz.
We have a simple set of guidelines for joining the NLANR cache hierarchy.
The NLANR hierarchy can provide you with an initial source for parent or sibling caches. Joining the NLANR global cache system will frequently improve the performance of your caching service.
Just enable these options in your squid.conf and you'll be registered:
cache_announce 24 announce_to sd.cache.nlanr.net:3131
NOTE: announcing your cache is not the same thing as joining the NLANR cache hierarchy. You can join the NLANR cache hierarchy without registering, and you can register without joining the NLANR cache hierarchy.
Visit the NLANR cache registration database to discover other caches near you. Keep in mind that just because a cache is registered in the database does not mean they are willing to be your parent/sibling/child. But it can't hurt to ask...
This entry has been moved to a different section.
Note: The information here is current for version 2.2.
If you are behind a firewall then you can't make direct connections to the outside world, so you must use a parent cache. Squid doesn't use ICP queries for a request if it's behind a firewall or if there is only one parent.
You can use the never_direct access list in
squid.conf to specify which requests must be forwarded to
your parent cache outside the firewall, and the always_direct access list
to specify which requests must not be forwarded. For example, if Squid
must connect directly to all servers that end with mydomain.com, but
must use the parent for all others, you would write:
acl INSIDE dstdomain .mydomain.com always_direct allow INSIDE never_direct allow all
You could also specify internal servers by IP address
acl INSIDE_IP dst 1.2.3.0/24 always_direct allow INSIDE_IP never_direct allow allNote, however that when you use IP addresses, Squid must perform a DNS lookup to convert URL hostnames to an address. Your internal DNS servers may not be able to lookup external domains.
If you use never_direct and you have multiple parent caches, then you probably will want to mark one of them as a default choice in case Squid can't decide which one to use. That is done with the default keyword on a cache_peer line. For example:
cache_peer xyz.mydomain.com parent 3128 0 default
Note: The information here is current for version 2.2.
First, you need to give Squid a parent cache. Second, you need to tell Squid it can not connect directly to origin servers. This is done with three configuration file lines:
cache_peer parentcache.foo.com parent 3128 0 no-query default acl all src 0.0.0.0/0.0.0.0 never_direct allow allNote, with this configuration, if the parent cache fails or becomes unreachable, then every request will result in an error message.
In case you want to be able to use direct connections when all the parents go down you should use a different approach:
cache_peer parentcache.foo.com parent 3128 0 no-query prefer_direct offThe default behaviour of Squid in the absence of positive ICP, HTCP, etc replies is to connect to the origin server instead of using parents. The prefer_direct off directive tells Squid to try parents first.
The dnsserver processes are used by squid because the gethostbyname(3) library routines used to
convert web sites names to their internet addresses
blocks until the function returns (i.e., the process that calls
it has to wait for a reply). Since there is only one squid
process, everyone who uses the cache would have to wait each
time the routine was called. This is why the dnsserver is
a separate process, so that these processes can block,
without causing blocking in squid.
It's very important that there are enough dnsserver processes to cope with every access you will need, otherwise squid will stop occasionally. A good rule of thumb is to make sure you have at least the maximum number of dnsservers squid has ever needed on your system, and probably add two to be on the safe side. In other words, if you have only ever seen at most three dnsserver processes in use, make at least five. Remember that a dnsserver is small and, if unused, will be swapped out.
First, find out if you have enough dnsserver processes running by looking at the Cachemanager dns output. Ideally, you should see that the first dnsserver handles a lot of requests, the second one less than the first, etc. The last dnsserver should have serviced relatively few requests. If there is not an obvious decreasing trend, then you need to increase the number of dns_children in the configuration file. If the last dnsserver has zero requests, then you definately have enough.
Another factor which affects the dnsserver service time is the proximity of your DNS resolver. Normally we do not recommend running Squid and named on the same host. Instead you should try use a DNS resolver (named) on a different host, but on the same LAN. If your DNS traffic must pass through one or more routers, this could be causing unnecessary delays.
Before you run the configure script, simply set the CACHE_HTTP_PORT environment variable.
setenv CACHE_HTTP_PORT 8080 ./configure make make install
With Squid-1.1 it is NOT possible. Each cache_dir is assumed to be the same size. The cache_swap setting defines the size of all cache_dir's taken together. If you have N cache_dir's then each one will hold cache_swap ÷ N Megabytes.
Most people have a disk partition dedicated to the Squid cache. You don't want to use the entire partition size. You have to leave some extra room. Currently, Squid is not very tolerant of running out of disk space.
Lets say you have a 9GB disk. Remember that disk manufacturers lie about the space available. A so-called 9GB disk usually results in about 8.5GB of raw, usable space. First, put a filesystem on it, and mount it. Then check the ``available space'' with your df program. Note that you lose some disk space to filesystem overheads, like superblocks, inodes, and directory entries. Also note that Unix normally keeps 10% free for itself. So with a 9GB disk, you're probably down to about 8GB after formatting.
Next, I suggest taking off another 10% or so for Squid overheads, and a "safe buffer." Squid normally puts its swap.state files in each cache directory. These grow in size until you rotate the logs, or restart squid. Also note that Squid performs better when there is more free space. So if performance is important to you, then take off even more space. Typically, for a 9GB disk, I recommend a cache_dir setting of 6000 to 7500 Megabytes:
cache_dir ... 7000 16 256
Its better to start out conservative. After the cache becomes full, look at the disk usage. If you think there is plenty of unused space, then increase the cache_dir setting a little.
If you're getting ``disk full'' write errors, then you definately need to decrease your cache size.
With Squid-1.1, yes, you will lose your cache. This is because version 1.1 uses a simplistic algorithm to distribute files between cache directories.
With Squid-2, you will not lose your existing cache. You can add and delete cache_dir's without affecting any of the others.
Several people on both the fwtk-users and the squid-users mailing asked about using Squid in combination with http-gw from the TIS toolkit. The most elegant way in my opinion is to run an internal Squid caching proxyserver which handles client requests and let this server forward it's requests to the http-gw running on the firewall. Cache hits won't need to be handled by the firewall.
In this example Squid runs on the same server as the http-gw, Squid uses 8000 and http-gw uses 8080 (web). The local domain is home.nl.
Either run http-gw as a daemon from the /etc/rc.d/rc.local (Linux Slackware):
exec /usr/local/fwtk/http-gw -daemon 8080or run it from inetd like this:
web stream tcp nowait.100 root /usr/local/fwtk/http-gw http-gwI increased the watermark to 100 because a lot of people run into problems with the default value.
Make sure you have at least the following line in /usr/local/etc/netperm-table:
http-gw: hosts 127.0.0.1You could add the IP-address of your own workstation to this rule and make sure the http-gw by itself works, like:
http-gw: hosts 127.0.0.1 10.0.0.1
The following settings are important:
http_port 8000 icp_port 0 cache_peer localhost.home.nl parent 8080 0 default acl HOME dstdomain .home.nl alwayws_direct allow HOME never_direct allow allThis tells Squid to use the parent for all domains other than home.nl. Below, access.log entries show what happens if you do a reload on the Squid-homepage:
872739961.631 1566 10.0.0.21 ERR_CLIENT_ABORT/304 83 GET http://www.squid-cache.org/ - DEFAULT_PARENT/localhost.home.nl - 872739962.976 1266 10.0.0.21 TCP_CLIENT_REFRESH/304 88 GET http://www.nlanr.net/Images/cache_now.gif - DEFAULT_PARENT/localhost.home.nl - 872739963.007 1299 10.0.0.21 ERR_CLIENT_ABORT/304 83 GET http://www.squid-cache.org/Icons/squidnow.gif - DEFAULT_PARENT/localhost.home.nl - 872739963.061 1354 10.0.0.21 TCP_CLIENT_REFRESH/304 83 GET http://www.squid-cache.org/Icons/Squidlogo2.gif - DEFAULT_PARENT/localhost.home.nl
http-gw entries in syslog:
Aug 28 02:46:00 memo http-gw[2052]: permit host=localhost/127.0.0.1 use of gateway (V2.0beta) Aug 28 02:46:00 memo http-gw[2052]: log host=localhost/127.0.0.1 protocol=HTTP cmd=dir dest=www.squid-cache.org path=/ Aug 28 02:46:01 memo http-gw[2052]: exit host=localhost/127.0.0.1 cmds=1 in=0 out=0 user=unauth duration=1 Aug 28 02:46:01 memo http-gw[2053]: permit host=localhost/127.0.0.1 use of gateway (V2.0beta) Aug 28 02:46:01 memo http-gw[2053]: log host=localhost/127.0.0.1 protocol=HTTP cmd=get dest=www.squid-cache.org path=/Icons/Squidlogo2.gif Aug 28 02:46:01 memo http-gw[2054]: permit host=localhost/127.0.0.1 use of gateway (V2.0beta) Aug 28 02:46:01 memo http-gw[2054]: log host=localhost/127.0.0.1 protocol=HTTP cmd=get dest=www.squid-cache.org path=/Icons/squidnow.gif Aug 28 02:46:01 memo http-gw[2055]: permit host=localhost/127.0.0.1 use of gateway (V2.0beta) Aug 28 02:46:01 memo http-gw[2055]: log host=localhost/127.0.0.1 protocol=HTTP cmd=get dest=www.nlanr.net path=/Images/cache_now.gif Aug 28 02:46:02 memo http-gw[2055]: exit host=localhost/127.0.0.1 cmds=1 in=0 out=0 user=unauth duration=1 Aug 28 02:46:03 memo http-gw[2053]: exit host=localhost/127.0.0.1 cmds=1 in=0 out=0 user=unauth duration=2 Aug 28 02:46:04 memo http-gw[2054]: exit host=localhost/127.0.0.1 cmds=1 in=0 out=0 user=unauth duration=3
To summarize:
Advantages:
Disadvantages:
-- Rodney van den Oever
When a proxy-cache is used, a server does not see the connection coming from the originating client. Many people like to implement access controls based on the client address. To accommodate these people, Squid adds its own request header called "X-Forwarded-For" which looks like this:
X-Forwarded-For: 128.138.243.150, unknown, 192.52.106.30Entries are always IP addresses, or the word unknown if the address could not be determined or if it has been disabled with the forwarded_for configuration option.
We must note that access controls based on this header are extremely weak and simple to fake. Anyone may hand-enter a request with any IP address whatsoever. This is perhaps the reason why client IP addresses have been omitted from the HTTP/1.1 specification.
Because of the weakness of this header, support for access controls based on X-Forwarder-For is not yet available in any officially released version of squid. However, unofficial patches are available from the follow_xff Squid development project and may be integrated into later versions of Squid once a suitable trust model have been developed.
Yes it can, however the way of doing it has changed from earlier versions of squid. As of squid-2.2 a more customisable method has been introduced. Please follow the instructions for the version of squid that you are using. As a default, no anonymizing is done.
If you choose to use the anonymizer you might wish to investigate the forwarded_for option to prevent the client address being disclosed. Failure to turn off the forwarded_for option will reduce the effectiveness of the anonymizer. Finally if you filter the User-Agent header using the fake_user_agent option can prevent some user problems as some sites require the User-Agent header.
With the introduction of squid 2.2 the anonoymizer has become more customisable. It now allows specification of exactly which headers will be allowed to pass. This is further extended in Squid-2.5 to allow headers to be anonymized conditionally.
For details see the documentation of the http_header_access and header_replace directives in squid.conf.default.
References: Anonymous WWW
Sure, just use the always_direct access list.
For example, if you want Squid to connect directly to hotmail.com servers, you can use these lines in your config file:
acl hotmail dstdomain .hotmail.com always_direct allow hotmail
Sure, there are few things you can do.
You can use the no_cache access list to make Squid never cache any response:
acl all src 0/0 no_cache deny all
With Squid-2.4 and later you can use the ``null'' storage module to avoid having a cache directory:
cache_dir null /tmp
Note: a null cache_dir does not disable caching, but it does save you from creating a cache structure if you have disabled caching with no_cache.
Note: the directory (e.g., /tmp) must exist so that squid can chdir to it, unless you also use the coredump_dir option.
To configure Squid for the ``null'' storage module, specify it on the configure command line:
./configure --enable-storeio=ufs,null ...
You can set the global reply_body_max_size parameter. This option controls the largest HTTP message body that will be sent to a cache client for one request.
If the HTTP response coming from the server has a Content-length
header, then Squid compares the content-length value to the
reply_body_max_size value. If the content-length is larger,
the server connection is closed and the user receives an error
message from Squid.
Some responses don't have Content-length
headers. In this case, Squid counts how many bytes are written
to the client. Once the limit is reached, the client's connection
is simply closed.
Note that ``creative'' user-agents will still be able to download really large files through the cache using HTTP/1.1 range requests.
Most web browsers available today support proxying and are easily configured to use a Squid server as a proxy. Some browsers support advanced features such as lists of domains or URL patterns that shouldn't be fetched through the proxy, or JavaScript automatic proxy configuration.
Select Network Preferences from the Options menu. On the Proxies page, click the radio button next to Manual Proxy Configuration and then click on the View button. For each protocol that your Squid server supports (by default, HTTP, FTP, and gopher) enter the Squid server's hostname or IP address and put the HTTP port number for the Squid server (by default, 3128) in the Port column. For any protocols that your Squid does not support, leave the fields blank.
Here is a screen shot of the Netscape Navigator manual proxy configuration screen.
Netscape Navigator's proxy configuration can be automated with JavaScript (for Navigator versions 2.0 or higher). Select Network Preferences from the Options menu. On the Proxies page, click the radio button next to Automatic Proxy Configuration and then fill in the URL for your JavaScript proxy configuration file in the text box. The box is too small, but the text will scroll to the right as you go.
Here is a screen shot of the Netscape Navigator automatic proxy configuration screen.
You may also wish to consult Netscape's documentation for the Navigator JavaScript proxy configuration
Here is a sample auto configuration JavaScript from Oskar Pearson:
//We (www.is.co.za) run a central cache for our customers that they
//access through a firewall - thus if they want to connect to their intranet
//system (or anything in their domain at all) they have to connect
//directly - hence all the "fiddling" to see if they are trying to connect
//to their local domain.
//Replace each occurrence of company.com with your domain name
//and if you have some kind of intranet system, make sure
//that you put it's name in place of "internal" below.
//We also assume that your cache is called "cache.company.com", and
//that it runs on port 8080. Change it down at the bottom.
//(C) Oskar Pearson and the Internet Solution (http://www.is.co.za)
function FindProxyForURL(url, host)
{
//If they have only specified a hostname, go directly.
if (isPlainHostName(host))
return "DIRECT";
//These connect directly if the machine they are trying to
//connect to starts with "intranet" - ie http://intranet
//Connect directly if it is intranet.*
//If you have another machine that you want them to
//access directly, replace "internal*" with that
//machine's name
if (shExpMatch( host, "intranet*")||
shExpMatch(host, "internal*"))
return "DIRECT";
//Connect directly to our domains (NB for Important News)
if (dnsDomainIs( host,"company.com")||
//If you have another domain that you wish to connect to
//directly, put it in here
dnsDomainIs(host,"sistercompany.com"))
return "DIRECT";
//So the error message "no such host" will appear through the
//normal Netscape box - less support queries :)
if (!isResolvable(host))
return "DIRECT";
//We only cache http, ftp and gopher
if (url.substring(0, 5) == "http:" ||
url.substring(0, 4) == "ftp:"||
url.substring(0, 7) == "gopher:")
//Change the ":8080" to the port that your cache
//runs on, and "cache.company.com" to the machine that
//you run the cache on
return "PROXY cache.company.com:8080; DIRECT";
//We don't cache WAIS
if (url.substring(0, 5) == "wais:")
return "DIRECT";
else
return "DIRECT";
}
For Mosaic and Lynx, you can set environment variables before starting the application. For example (assuming csh or tcsh):
% setenv http_proxy http://mycache.example.com:3128/ % setenv gopher_proxy http://mycache.example.com:3128/ % setenv ftp_proxy http://mycache.example.com:3128/
For Lynx you can also edit the lynx.cfg file to configure proxy usage. This has the added benefit of causing all Lynx users on a system to access the proxy without making environment variable changes for each user. For example:
http_proxy:http://mycache.example.com:3128/
ftp_proxy:http://mycache.example.com:3128/
gopher_proxy:http://mycache.example.com:3128/
There's one nasty side-effect to using auto-proxy scripts: if you start the web browser it will try and load the auto-proxy-script.
If your script isn't available either because the web server hosting the script is down or your workstation can't reach the web server (e.g. because you're working off-line with your notebook and just want to read a previously saved HTML-file) you'll get different errors depending on the browser you use.
The Netscape browser will just return an error after a timeout (after that it tries to find the site 'www.proxy.com' if the script you use is called 'proxy.pac').
The Microsoft Internet Explorer on the other hand won't even start, no window displays, only after about 1 minute it'll display a window asking you to go on with/without proxy configuration.
The point is that your workstations always need to locate the proxy-script. I created some extra redundancy by hosting the script on two web servers (actually Apache web servers on the proxy servers themselves) and adding the following records to my primary nameserver:
proxy CNAME proxy1
CNAME proxy2
The clients just refer to 'http://proxy/proxy.pac'. This script looks like this:
function FindProxyForURL(url,host)
{
// Hostname without domainname or host within our own domain?
// Try them directly:
// http://www.domain.com actually lives before the firewall, so
// make an exception:
if ((isPlainHostName(host)||dnsDomainIs( host,".domain.com")) &&
!localHostOrDomainIs(host, "www.domain.com"))
return "DIRECT";
// First try proxy1 then proxy2. One server mostly caches '.com'
// to make sure both servers are not
// caching the same data in the normal situation. The other
// server caches the other domains normally.
// If one of 'm is down the client will try the other server.
else if (shExpMatch(host, "*.com"))
return "PROXY proxy1.domain.com:8080; PROXY proxy2.domain.com:8081; DIRECT";
return "PROXY proxy2.domain.com:8081; PROXY proxy1.domain.com:8080; DIRECT";
}
I made sure every client domain has the appropriate 'proxy' entry. The clients are automatically configured with two nameservers using DHCP.
-- Rodney van den Oever
The Sharp Super Proxy Script page contains a lot of good information about hash-based proxy auto-configuration scripts. With these you can distribute the load between a number of caching proxies.
Select Options from the View menu. Click on the Connection tab. Tick the Connect through Proxy Server option and hit the Proxy Settings button. For each protocol that your Squid server supports (by default, HTTP, FTP, and gopher) enter the Squid server's hostname or IP address and put the HTTP port number for the Squid server (by default, 3128) in the Port column. For any protocols that your Squid does not support, leave the fields blank.
Here is a screen shot of the Internet Explorer proxy configuration screen.
Microsoft is also starting to support Netscape-style JavaScript automated proxy configuration. As of now, only MSIE version 3.0a for Windows 3.1 and Windows NT 3.51 supports this feature (i.e., as of version 3.01 build 1225 for Windows 95 and NT 4.0, the feature was not included).
If you have a version of MSIE that does have this feature, elect Options from the View menu. Click on the Advanced tab. In the lower left-hand corner, click on the Automatic Configuration button. Fill in the URL for your JavaScript file in the dialog box it presents you. Then exit MSIE and restart it for the changes to take effect. MSIE will reload the JavaScript file every time it starts.
Netmanage WebSurfer supports manual proxy configuration and exclusion lists for hosts or domains that should not be fetched via proxy (this information is current as of WebSurfer 5.0). Select Preferences from the Settings menu. Click on the Proxies tab. Select the Use Proxy options for HTTP, FTP, and gopher. For each protocol that enter the Squid server's hostname or IP address and put the HTTP port number for the Squid server (by default, 3128) in the Port boxes. For any protocols that your Squid does not support, leave the fields blank.
Take a look at this screen shot if the instructions confused you.
On the same configuration window, you'll find a button to bring up the exclusion list dialog box, which will let you enter some hosts or domains that you don't want fetched via proxy. It should be self-explanatory, but you might look at this screen shot just for fun anyway.
Select Proxy Servers... from the Preferences menu. Check each protocol that your Squid server supports (by default, HTTP, FTP, and Gopher) and enter the Squid server's address as hostname:port (e.g. mycache.example.com:3128 or 123.45.67.89:3128). Click on Okay to accept the setup.
Notes:
-- Hume Smith
Insert your username in the host part of the URL, for example:
ftp://joecool@ftp.foo.org/Squid should then prompt you for your account password. Alternatively, you can specify both your username and password in the URL itself:
ftp://joecool:secret@ftp.foo.org/However, we certainly do not recommend this, as it could be very easy for someone to see or grab your password.
You may like to start by reading the Expired Internet-Draft that describes WPAD.
After reading the 8 steps below, if you don't understand any of the terms or methods mentioned, you probably shouldn't be doing this. Implementing wpad requires you to fully understand:
This is not a recommendation for any product or version. As far as I know IE5 is the only browser out now implementing wpad. I think wpad is an excellent feature that will return several hours of life per month. Hopefully, all browser clients will implement it as well. But it will take years for all the older browsers to fade away though.
I have only focused on the domain name method, to the exclusion of the DHCP method. I think the dns method might be easier for most people. I don't currently, and may never, fully understand wpad and IE5, but this method worked for me. It may work for you.
But if you'd rather just have a go ...
Andrei Ivanov notes that you should be able to use an HTTP redirect if you want to store the wpad.dat file somewhere else. You can probably even redirect wpad.dat to proxy.pac:
Redirect /wpad.dat http://racoon.riga.lv/proxy.pac
http://www.your.domain.name/wpad.dat should bring up
the script text in your browser window.
application/x-ns-proxy-autoconfig datAnd then restart your web server, for new mime type to work.
http://www.your.domain.name/wpad.dat Test that
that all works as per your script and network. There's no point
continuing until this works ...
wpad.your.domain.name resolves to the host above where
you have a functioning auto config script running. You should
now be able to use http://wpad.your.domain.name/wpad.dat
as the Auto Config Script location in step 5 above.
Anybody using these steps to install and test, please feel free to make notes, corrections or additions for improvements, and post back to the squid list...
There are probably many more tricks and tips which hopefully will be detailed here in the future. Things like wpad.dat files being served from the proxy server themselves, maybe with a round robin dns setup for the WPAD host.
You can also use DHCP to configure browsers for WPAD. This technique allows you to set any URL as the PAC URL. For ISC DHCPD, enter a line like this in your dhcpd.conf file:
option wpad code 252 = text; option wpad "http://www.example.com/proxy.pac";
Replace the hostname with the name or address of your own server.
Ilja Pavkovic notes that the DHCP mode does not work reliably with every version of Internet Explorer. The DNS name method to find wpad.dat is more reliable.
Another user adds that IE 6.01 seems to strip the last character from the URL. By adding a trailing newline, he is able to make it work with both IE 5.0 and 6.0:<
option wpad "http://www.example.com/proxy.pac\n";
There was a bug in the 5.0x releases of Internet Explorer in which IE cropped any trailing slash off an FTP URL. The URL showed up correctly in the browser's ``Address:'' field, however squid logs show that the trailing slash was being taken off.
An example of where this impacted squid if you had a setup where squid would go direct for FTP directory listings but forward a request to a parent for FTP file transfers. This was useful if your upstream proxy was an older version of Squid or another vendors software which displayed directory listings with broken icons and you wanted your own local version of squid to generate proper FTP directory listings instead. The workaround for this is to add a double slash to any directory listing in which the slash was important, or else upgrade to IE 5.5. (Or use Netscape)
When using authentication with Internet Explorer 6 SP1, you may encounter issues when you first launch Internet Explorer. The problem will show itself when you first authenticate, you will receive a "Page Cannot Be Displayed" error. However, if you click refresh, the page will be correctly displayed.
This only happens immediately after you authenticate.
This is not a Squid error or bug. Microsoft broke the Basic Authentication when they put out IE6 SP1.
There is a knowledgebase article ( KB 331906) regarding this issue, which contains a link to a downloadable "hot fix." They do warn that this code is not "regression tested" but so far there have not been any reports of this breaking anything else. The problematic file is wininet.dll. Please note that this hotfix is included in the latest security update.
Lloyd Parkes notes that the article references another article, KB 312176. He says that you must not have the registry entry that KB 312176 encourages users to add to their registry.
According to Joao Coutinho, this simple solution also corrects the problem:
Another possible workaround to these problems is to make the ERR_CACHE_ACCESS_DENIED larger than 1460 bytes. This should trigger IE to handle the authentication in a slightly different manner.
The logs are a valuable source of information about Squid workloads and performance. The logs record not only access information, but also system configuration errors and resource consumption (eg, memory, disk space). There are several log file maintained by Squid. Some have to be explicitely activated during compile time, others can safely be deactivated during run-time.
There are a few basic points common to all log files. The time stamps logged into the log files are usually UTC seconds unless stated otherwise. The initial time stamp usually contains a millisecond extension.
If you run your Squid from the RunCache script, a file squid.out contains the Squid startup times, and also all fatal errors, e.g. as produced by an assert() failure. If you are not using RunCache, you will not see such a file.
The cache.log file contains the debug and error messages that Squid generates. If you start your Squid using the default RunCache script, or start it with the -s command line option, a copy of certain messages will go into your syslog facilities. It is a matter of personal preferences to use a separate file for the squid log data.
From the area of automatic log file analysis, the cache.log file does not have much to offer. You will usually look into this file for automated error reports, when programming Squid, testing new features, or searching for reasons of a perceived misbehaviour, etc.
The user agent log file is only maintained, if
From the user agent log file you are able to find out about distributation of browsers of your clients. Using this option in conjunction with a loaded production squid might not be the best of all ideas.
The store.log file covers the objects currently kept on disk or removed ones. As a kind of transaction log it is ususally used for debugging purposes. A definitive statement, whether an object resides on your disks is only possible after analysing the complete log file. The release (deletion) of an object may be logged at a later time than the swap out (save to disk).
The store.log file may be of interest to log file analysis which looks into the objects on your disks and the time they spend there, or how many times a hot object was accessed. The latter may be covered by another log file, too. With knowledge of the cache_dir configuration option, this log file allows for a URL to filename mapping without recursing your cache disks. However, the Squid developers recommend to treat store.log primarily as a debug file, and so should you, unless you know what you are doing.
The print format for a store log entry (one line) consists of eleven space-separated columns, compare with the storeLog() function in file src/store_log.c:
"%9d.%03d %-7s %02d %08X %4d %9d %9d %9d %s %d/%d %s %s\n"
The timestamp when the line was logged in UTC with a millisecond fraction.
The action the object was sumitted to, compare with src/store_log.c:
The cache_dir number this object was stored into, starting at 0 for your first cache_dir line.
The file number for the object storage file. Please note that the path to this file is calculated according to your cache_dir configuration.
A file number of FFFFFFFF denominates "memory only" objects. Any action code for such a file number refers to an object which existed only in memory, not on disk. For instance, if a RELEASE code was logged with file number FFFFFFFF, the object existed only in memory, and was released from memory.
The HTTP reply status code.
The value of the HTTP "Date: " reply header.
The value of the HTTP "Last-Modified: " reply header.
The value of the HTTP "Expires: " reply header.
The HTTP "Content-Type" major value, or "unknown" if it cannot be determined.
This column consists of two slash separated fields:
If the advertised (or expected) length is missing, it will be set to zero. If the advertised length is not zero, but not equal to the real length, the object will be realeased from the cache.
The request method for the object, e.g. GET.
The key to the object, usually the URL.
The timestamp format for the columns Date to Expires are all expressed in UTC seconds. The actual values are parsed from the HTTP reply headers. An unparsable header is represented by a value of -1, and a missing header is represented by a value of -2.
The column key usually contains just the URL of the object. Some objects though will never become public. Thus the key is said to include a unique integer number and the request method in addition to the URL.
This logfile exists for Squid-1.0 only. The format is
[date] URL peerstatus peerhost
Most log file analysis program are based on the entries in access.log. Currently, there are two file formats possible for the log file, depending on your configuration for the emulate_httpd_log option. By default, Squid will log in its native log file format. If the above option is enabled, Squid will log in the common log file format as defined by the CERN web daemon.
The common log file format contains other information than the native log file, and less. The native format contains more information for the admin interested in cache evaluation.
The Common Logfile Format is used by numerous HTTP servers. This format consists of the following seven fields:
remotehost rfc931 authuser [date] "method URL" status bytes
It is parsable by a variety of tools. The common format contains different information than the native log file format. The HTTP version is logged, which is not logged in native log file format.
The native format is different for different major versions of Squid. For Squid-1.0 it is:
time elapsed remotehost code/status/peerstatus bytes method URL
For Squid-1.1, the information from the hierarchy.log was moved into access.log. The format is:
time elapsed remotehost code/status bytes method URL rfc931 peerstatus/peerhost type
For Squid-2 the columns stay the same, though the content within may change a little.
The native log file format logs more and different information than the common log file format: the request duration, some timeout information, the next upstream server address, and the content type.
There exist tools, which convert one file format into the other. Please mind that even though the log formats share most information, both formats contain information which is not part of the other format, and thus this part of the information is lost when converting. Especially converting back and forth is not possible without loss.
squid2common.pl is a conversion utility, which converts any of the squid log file formats into the old CERN proxy style output. There exist tools to analyse, evaluate and graph results from that format.
It is recommended though to use Squid's native log format due to its greater amount of information made available for later analysis. The print format line for native access.log entries looks like this:
"%9d.%03d %6d %s %s/%03d %d %s %s %s %s%s/%s %s"
Therefore, an access.log entry usually consists of (at least) 10 columns separated by one ore more spaces:
A Unix timestamp as UTC seconds with a millisecond resolution. You can convert Unix timestamps into something more human readable using this short perl script:
#! /usr/bin/perl -p s/^\d+\.\d+/localtime $&/e;
The elapsed time considers how many milliseconds the transaction busied the cache. It differs in interpretation between TCP and UDP:
Please note that the entries are logged after the reply finished being sent, not during the lifetime of the transaction.
The IP address of the requesting instance, the client IP address. The client_netmask configuration option can distort the clients for data protection reasons, but it makes analysis more difficult. Often it is better to use one of the log file anonymizers.
Also, the log_fqdn configuration option may log the fully qualified domain name of the client instead of the dotted quad. The use of that option is discouraged due to its performance impact.
This column is made up of two entries separated by a slash. This column encodes the transaction result:
Several codes from older versions are no longer available, were renamed, or split. Especially the ERR_ codes do not seem to appear in the log file any more. Also refer to section Squid result codes for details on the codes no longer available in Squid-2.
The NOVM versions and Squid-2 also rely on the Unix buffer cache, thus you will see less TCP_MEM_HITs than with a Squid-1. Basically, the NOVM feature relies on read() to obtain an object, but due to the kernel buffer cache, no disk activity is needed. Only small objects (below 8KByte) are kept in Squid's part of main memory.
The size is the amount of data delivered to the client. Mind that this does not constitute the net object size, as headers are also counted. Also, failed requests may deliver an error page, the size of which is also logged here.
The request method to obtain an object. Please refer to section request-methods for available methods. If you turned off log_icp_queries in your configuration, you will not see (and thus unable to analyse) ICP exchanges. The PURGE method is only available, if you have an ACL for ``method purge'' enabled in your configuration file.
This column contains the URL requested. Please note that the log file may contain whitespaces for the URI. The default configuration for uri_whitespace denies whitespaces, though.
The eigth column may contain the ident lookups for the requesting client. Since ident lookups have performance impact, the default configuration turns ident_loookups off. If turned off, or no ident information is available, a ``-'' will be logged.
The hierarchy information consists of three items:
The content type of the object as seen in the HTTP reply header. Please note that ICP exchanges usually don't have any content type, and thus are logged ``-''. Also, some weird replies have content types ``:'' or even empty ones.
There may be two more columns in the access.log, if the (debug) option log_mime_headers is enabled In this case, the HTTP request headers are logged between a ``['' and a ``]'', and the HTTP reply headers are also logged between ``['' and ``]''. All control characters like CR and LF are URL-escaped, but spaces are not escaped! Parsers should watch out for this.
The TCP_ codes refer to requests on the HTTP port (usually 3128). The UDP_ codes refer to requests on the ICP port (usually 3130). If ICP logging was disabled using the log_icp_queries option, no ICP replies will be logged.
The following result codes were taken from a Squid-2, compare with the log_tags struct in src/access_log.c:
A valid copy of the requested object was in the cache.
The requested object was not in the cache.
The requested object was cached but STALE. The IMS query for the object resulted in "304 not modified".
The requested object was cached but STALE. The IMS query failed and the stale object was delivered.
The requested object was cached but STALE. The IMS query returned the new content.
The client issued a "no-cache" pragma, or some analogous cache control command along with the request. Thus, the cache has to refetch the object.
The client issued an IMS request for an object which was in the cache and fresh.
The object was believed to be in the cache, but could not be accessed.
Request for a negatively cached object, e.g. "404 not found", for which the cache believes to know that it is inaccessible. Also refer to the explainations for negative_ttl in your squid.conf file.
A valid copy of the requested object was in the cache and it was in memory, thus avoiding disk accesses.
Access was denied for this request.
The requested object was retrieved from the cache during offline mode. The offline mode never validates any object, see offline_mode in squid.conf file.
A valid copy of the requested object was in the cache.
The requested object is not in this cache.
Access was denied for this request.
An invalid request was received.
During "-Y" startup, or during frequent failures, a cache in hit only mode will return either UDP_HIT or this code. Neighbours will thus only fetch hits.
Seen with errors and cachemgr requests.
The following codes are no longer available in Squid-2:
Errors are now contained in the status code.
See: TCP_CLIENT_REFRESH_MISS.
See: TCP_SWAPFAIL_MISS.
Deleted, TCP_IMS_HIT used instead.
Hit objects are no longer available.
See: UDP_MISS_NOFETCH.
These are taken from RFC 2616 and verified for Squid. Squid-2 uses almost all codes except 307 (Temporary Redirect), 416 (Request Range Not Satisfiable), and 417 (Expectation Failed). Extra codes include 0 for a result code being unavailable, and 600 to signal an invalid header, a proxy error. Also, some definitions were added as for RFC 2518 (WebDAV). Yes, there are really two entries for status code 424, compare with http_status in src/enums.h:
000 Used mostly with UDP traffic. 100 Continue 101 Switching Protocols *102 Processing 200 OK 201 Created 202 Accepted 203 Non-Authoritative Information 204 No Content 205 Reset Content 206 Partial Content *207 Multi Status 300 Multiple Choices 301 Moved Permanently 302 Moved Temporarily 303 See Other 304 Not Modified 305 Use Proxy [307 Temporary Redirect] 400 Bad Request 401 Unauthorized 402 Payment Required 403 Forbidden 404 Not Found 405 Method Not Allowed 406 Not Acceptable 407 Proxy Authentication Required 408 Request Timeout 409 Conflict 410 Gone 411 Length Required 412 Precondition Failed 413 Request Entity Too Large 414 Request URI Too Large 415 Unsupported Media Type [416 Request Range Not Satisfiable] [417 Expectation Failed] *424 Locked *424 Failed Dependency *433 Unprocessable Entity 500 Internal Server Error 501 Not Implemented 502 Bad Gateway 503 Service Unavailable 504 Gateway Timeout 505 HTTP Version Not Supported *507 Insufficient Storage 600 Squid header parsing error
Squid recognizes several request methods as defined in RFC 2616. Newer versions of Squid (2.2.STABLE5 and above) also recognize RFC 2518 ``HTTP Extensions for Distributed Authoring -- WEBDAV'' extensions.
method defined cachabil. meaning --------- ---------- ---------- ------------------------------------------- GET HTTP/0.9 possibly object retrieval and simple searches. HEAD HTTP/1.0 possibly metadata retrieval. POST HTTP/1.0 CC or Exp. submit data (to a program). PUT HTTP/1.1 never upload data (e.g. to a file). DELETE HTTP/1.1 never remove resource (e.g. file). TRACE HTTP/1.1 never appl. layer trace of request route. OPTIONS HTTP/1.1 never request available comm. options. CONNECT HTTP/1.1r3 never tunnel SSL connection. ICP_QUERY Squid never used for ICP based exchanges. PURGE Squid never remove object from cache. PROPFIND rfc2518 ? retrieve properties of an object. PROPATCH rfc2518 ? change properties of an object. MKCOL rfc2518 never create a new collection. COPY rfc2518 never create a duplicate of src in dst. MOVE rfc2518 never atomically move src to dst. LOCK rfc2518 never lock an object against modifications. UNLOCK rfc2518 never unlock an object.
The following hierarchy codes are used with Squid-2:
For TCP HIT, TCP failures, cachemgr requests and all UDP requests, there is no hierarchy information.
The object was fetched from the origin server.
The object was fetched from a sibling cache which replied with UDP_HIT.
The object was requested from a parent cache which replied with UDP_HIT.
No ICP queries were sent. This parent was chosen because it was marked ``default'' in the config file.
The object was requested from the only parent appropriate for the given URL.
The object was fetched from the first parent in the list of parents.
The object was fetched from the origin server, because no parents existed for the given URL.
The object was fetched from the parent with the fastest (possibly weighted) round trip time.
This parent was chosen, because it included the the lowest RTT measurement to the origin server. See also the closests-only peer configuration option.
The parent selection was based on our own RTT measurements.
Our own RTT measurements returned a shorter time than any parent.
The object could not be requested because of a firewall configuration, see also never_direct and related material, and no parents were available.
The origin site was chosen, because the source ping arrived fastest.
No ICP replies were received from any parent. The parent was chosen, because it was marked for round robin in the config file and had the lowest usage count.
The peer was chosen, because the cache digest predicted a hit. This option was later replaced in order to distinguish between parents and siblings.
The parent was chosen, because the cache digest predicted a hit.
The sibling was chosen, because the cache digest predicted a hit.
This output seems to be unused?
The peer was selected by CARP.
part of src/peer_select.c:hier_strings[].
part of src/peer_select.c:hier_strings[].
Almost any of these may be preceded by 'TIMEOUT_' if the two-second (default) timeout occurs waiting for all ICP replies to arrive from neighbors, see also the icp_query_timeout configuration option.
The following hierarchy codes were removed from Squid-2:
code meaning -------------------- ------------------------------------------------- PARENT_UDP_HIT_OBJ hit objects are not longer available. SIBLING_UDP_HIT_OBJ hit objects are not longer available. SSL_PARENT_MISS SSL can now be handled by squid. FIREWALL_IP_DIRECT No special logging for hosts inside the firewall. LOCAL_IP_DIRECT No special logging for local networks.
This file has a rather unfortunate name. It also is often called the swap log. It is a record of every cache object written to disk. It is read when Squid starts up to ``reload'' the cache. If you remove this file when squid is NOT running, you will effectively wipe out your cache contents. If you remove this file while squid IS running, you can easily recreate it. The safest way is to simply shutdown the running process:
% squid -k shutdownThis will disrupt service, but at least you will have your swap log back. Alternatively, you can tell squid to rotate its log files. This also causes a clean swap log to be written.
% squid -k rotate
For Squid-1.1, there are six fields:
In Squid-2, the swap log file is now called swap.state. This is a binary file that includes MD5 checksums, and StoreEntry fields. Please see the Programmers Guide for information on the contents and format of that file.
If you remove swap.state while Squid is running, simply send Squid the signal to rotate its log files:
% squid -k rotateAlternatively, you can tell Squid to shutdown and it will rewrite this file before it exits.
If you remove the swap.state while Squid is not running, you will not lose your entire cache. In this case, Squid will scan all of the cache directories and read each swap file to rebuild the cache. This can take a very long time, so you'll have to be patient.
By default the swap.state file is stored in the top-level of each cache_dir. You can move the logs to a different location with the cache_swap_log option.
You should never delete access.log, store.log, cache.log, or swap.state while Squid is running. With Unix, you can delete a file when a process has the file opened. However, the filesystem space is not reclaimed until the process closes the file.
If you accidentally delete swap.state while Squid is running, you can recover it by following the instructions in the previous questions. If you delete the others while Squid is running, you can not recover them.
The correct way to maintain your log files is with Squid's ``rotate'' feature. You should rotate your log files at least once per day. The current log files are closed and then renamed with numeric extensions (.0, .1, etc). If you want to, you can write your own scripts to archive or remove the old log files. If not, Squid will only keep up to logfile_rotate versions of each log file. The logfile rotation procedure also writes a clean swap.state file, but it does not leave numbered versions of the old files.
If you set logfile_rotate to 0, Squid simply closes and then re-opens the logs. This allows third-party logfile management systems, such as newsyslog, to maintain the log files.
To rotate Squid's logs, simple use this command:
squid -k rotateFor example, use this cron entry to rotate the logs at midnight:
0 0 * * * /usr/local/squid/bin/squid -k rotate
For Squid 2.4:
To disable access.log:
cache_access_log /dev/null
To disable store.log:
cache_store_log none
To disable cache.log:
cache_log /dev/null
For Squid 2.5:
To disable access.log:
cache_access_log none
To disable store.log:
cache_store_log none
To disable cache.log:
cache_log /dev/null
Note : It is a bad idea to disable the cache.log because this file contains many important status and debugging messages. However, if you really want to, you can.
Warning : If /dev/null is specified to any of the above log files, logfile rotate must also be set to 0 or else risk Squid rotating away /dev/null making it a plain log file.
Tip : Instead of disabling the log files, it is advisable to use a smaller value for logfile_rotate and properly rotating Squid's log files in your cron. That way, your log files are more controllable and self-maintained by your system.
You need to rotate your log files with a cron job. For example:
0 0 * * * /usr/local/squid/bin/squid -k rotate
If you set logfile_rotate to 0, Squid simply closes and then re-opens the logs. This allows third-party logfile management systems, such as newsyslog, to maintain the log files.
The preferred log file for analysis is the access.log file in native format. For long term evaluations, the log file should be obtained at regular intervals. Squid offers an easy to use API for rotating log files, in order that they may be moved (or removed) without disturbing the cache operations in progress. The procedures were described above.
Depending on the disk space allocated for log file storage, it is recommended to set up a cron job which rotates the log files every 24, 12, or 8 hour. You will need to set your logfile_rotate to a sufficiently large number. During a time of some idleness, you can safely transfer the log files to your analysis host in one burst.
Before transport, the log files can be compressed during off-peak time. On the analysis host, the log file are concatinated into one file, so one file for 24 hours is the yield. Also note that with log_icp_queries enabled, you might have around 1 GB of uncompressed log information per day and busy cache. Look into you cache manager info page to make an educated guess on the size of your log files.
The EU project DESIRE developed some some basic rules to obey when handling and processing log files:
This message means that the requested object was in ``Delete Behind'' mode and the user aborted the transfer. An object will go into ``Delete Behind'' mode if
This means that a timeout occurred while the object was being transferred. Most likely the retrieval of this object was very slow (or it stalled before finishing) and the user aborted the request. However, depending on your settings for quick_abort, Squid may have continued to try retrieving the object. Squid imposes a maximum amount of time on all open sockets, so after some amount of time the stalled request was aborted and logged win an ERR_LIFETIME_EXP message.
I've been asked to retrieve an object which was accidentally destroyed at the source for recovery. So, how do I figure out where the things are so I can copy them out and strip off the headers?
The following method applies only to the Squid-1.1 versions:
Use grep to find the named object (Url) in the cache/log file. The first field in this file is an integer file number.
Then, find the file fileno-to-pathname.pl from the ``scripts'' directory of the Squid source distribution. The usage is
perl fileno-to-pathname.pl [-c squid.conf]file numbers are read on stdin, and pathnames are printed on stdout.
Sort of. You can use store.log to find out if a particular response was cached.
Cached responses are logged with the SWAPOUT tag. Uncached responses are logged with the RELEASE tag.
However, your analysis must also consider that when a cached response is removed from the cache (for example due to cache replacement) it is also logged in store.log with the RELEASE tag. To differentiate these two, you can look at the filenumber (3rd) field. When an uncachable response is released, the filenumber is FFFFFFFF (-1). Any other filenumber indicates a cached response was released.
The Squid distribution includes a CGI utility called cachemgr.cgi which can be used to view squid statistics with a web browser. This document has a section devoted to cachemgr.cgi usage which you should consult for more information.
sort -r -n +4 -5 access.log | awk '{print $5, $7}' | head -25
Note: The information here is current for version 2.2 and later.
First of all, you must stop Squid of course. You can use the command:
% squid -k shutdown
The fastest way to restart with an entirely clean cache is to over write the swap.state files for each cache_dir in your config file. Note, you can not just remove the swap.state file, or truncate it to zero size. Instead, you should put just one byte of garbage there. For example:
% echo "" > /cache1/swap.stateRepeat that for every cache_dir, then restart Squid. Be sure to leave the swap.state file with the same owner and permissions that it had before!
Another way, which takes longer, is to have squid recreate all the cache_dir directories. But first you must move the existing directories out of the way. For example, you can try this:
% cd /cache1 % mkdir JUNK % mv ?? swap.state* JUNK % rm -rf JUNK &Repeat this for your other cache_dir's, then tell Squid to create new directories:
% squid -z
by Rodney van den Oever, and James R Grinter
The RealPlayer (and RealPlayer Plus) manual states:
Use HTTP Only Select this option if you are behind a firewall and cannot receive data through TCP. All data will be streamed through HTTP. Note: You may not be able to receive some content if you select this option.
Again, from the documentation:
RealPlayer 4.0 identifies itself to the firewall when making a request for content to a RealServer. The following string is attached to any URL that the Player requests using HTTP GET: /SmpDsBhgRl Thus, to identify an HTTP GET request from the RealPlayer, look for: http://[^/]+/SmpDsBhgRl The Player can also be identified by the mime type in a POST to the RealServer. The RealPlayer POST has the following mime type: "application/x-pncmd"
Note that the first request is a POST, and the second has a '?' in the URL, so standard Squid configurations would treat it as non-cachable. It also looks rather ``magic.''
HTTP is an alternative delivery mechanism introduced with version 3 players, and it allows a reasonable approximation to ``streaming'' data - that is playing it as you receive it.
It isn't available in the general case: only if someone has made the realaudio file available via an HTTP server, or they're using a version 4 server, they've switched it on, and you're using a version 4 client. If someone has made the file available via their HTTP server, then it'll be cachable. Otherwise, it won't be (as far as we can tell.)
The more common RealAudio link connects via their own pnm: method and is transferred using their proprietary protocol (via TCP or UDP) and not using HTTP. It can't be cached nor proxied by Squid, and requires something such as the simple proxy that Progressive Networks themselves have made available, if you're in a firewall/no direct route situation. Their product does not cache (and I don't know of any software available that does.)
Some confusion arises because there is also a configuration option to use an
HTTP proxy (such as Squid) with the Realaudio/RealVideo players. This is
because the players can fetch the ``.ram'' file that contains the pnm:
reference for the audio/video stream. They fetch that .ram file from an HTTP
server, using HTTP.
Squid does not allow you to purge objects unless it is configured with access controls in squid.conf. First you must add something like
acl PURGE method PURGE acl localhost src 127.0.0.1 http_access allow PURGE localhost http_access deny PURGEThe above only allows purge requests which come from the local host and denies all other purge requests.
To purge an object, you can use the squidclient program:
squidclient -m PURGE http://www.miscreant.com/If the purge was successful, you will see a ``200 OK'' response:
HTTP/1.0 200 OK Date: Thu, 17 Jul 1997 16:03:32 GMT Server: Squid/1.1.14If the object was not found in the cache, you will see a ``404 Not Found'' response:
HTTP/1.0 404 Not Found Date: Thu, 17 Jul 1997 16:03:22 GMT Server: Squid/1.1.14
As of version 1.1.9, Squid is able to utilize ICMP Round-Trip-Time (RTT) measurements to select the optimal location to forward a cache miss. Previously, cache misses would be forwarded to the parent cache which returned the first ICP reply message. These were logged with FIRST_PARENT_MISS in the access.log file. Now we can select the parent which is closest (RTT-wise) to the origin server.
It is more important that your parent caches enable the ICMP features. If you are acting as a parent, then you may want to enable ICMP on your cache. Also, if your cache makes RTT measurements, it will fetch objects directly if your cache is closer than any of the parents.
If you want your Squid cache to measure RTT's to origin servers, Squid must be compiled with the USE_ICMP option. This is easily accomplished by uncommenting "-DUSE_ICMP=1" in src/Makefile and/or src/Makefile.in.
An external program called pinger is responsible for sending and receiving ICMP packets. It must run with root privileges. After Squid has been compiled, the pinger program must be installed separately. A special Makefile target will install pinger with appropriate permissions.
% make install % su # make install-pingerThere are three configuration file options for tuning the measurement database on your cache. netdb_low and netdb_high specify high and low water marks for keeping the database to a certain size (e.g. just like with the IP cache). The netdb_ttl option specifies the minimum rate for pinging a site. If netdb_ttl is set to 300 seconds (5 minutes) then an ICMP packet will not be sent to the same site more than once every five minutes. Note that a site is only pinged when an HTTP request for the site is received.
Another option, minimum_direct_hops can be used to try finding servers which are close to your cache. If the measured hop count to the origin server is less than or equal to minimum_direct_hops, the request will be forwarded directly to the origin server.
Your parent caches can be asked to include the RTT measurements in their ICP replies. To do this, you must enable query_icmp in your config file:
query_icmp onThis causes a flag to be set in your outgoing ICP queries.
If your parent caches return ICMP RTT measurements then the eighth column of your access.log will have lines similar to:
CLOSEST_PARENT_MISS/it.cache.nlanr.netIn this case, it means that it.cache.nlanr.net returned the lowest RTT to the origin server. If your cache measured a lower RTT than any of the parents, the request will be logged with
CLOSEST_DIRECT/www.sample.com
The measurement database can be viewed from the cachemgr by selecting "Network Probe Database." Hostnames are aggregated into /24 networks. All measurements made are averaged over time. Measurements are made to specific hosts, taken from the URLs of HTTP requests. The recv and sent fields are the number of ICMP packets sent and received. At this time they are only informational.
A typical database entry looks something like this:
Network recv/sent RTT Hops Hostnames
192.41.10.0 20/ 21 82.3 6.0 www.jisedu.org www.dozo.com
bo.cache.nlanr.net 42.0 7.0
uc.cache.nlanr.net 48.0 10.0
pb.cache.nlanr.net 55.0 10.0
it.cache.nlanr.net 185.0 13.0
This means we have sent 21 pings to both www.jisedu.org and
www.dozo.com. The average RTT is 82.3 milliseconds. The
next four lines show the measured values from our parent
caches. Since bo.cache.nlanr.net has the lowest RTT,
it would be selected as the location to forward a request
for a www.jisedu.org or www.dozo.com URL.
When Squid receives an If-Modified-Since request, it will not forward the request unless the object needs to be refreshed according to the refresh_pattern rules. If the request does need to be refreshed, then it will be logged as TCP_REFRESH_HIT or TCP_REFRESH_MISS.
If the request is not forwarded, Squid replies to the IMS request according to the object in its cache. If the modification times are the same, then Squid returns TCP_IMS_HIT. If the modification times are different, then Squid returns TCP_IMS_MISS. In most cases, the cached object will not have changed, so the result is TCP_IMS_HIT. Squid will only return TCP_IMS_MISS if some other client causes a newer version of the object to be pulled into the cache.
In Squid-2, you use the no_cache option to specify uncachable requests. For example, this makes all responses from origin servers in the 10.0.1.0/24 network uncachable:
acl Local dst 10.0.1.0/24 no_cache deny Local
This example makes all URL's with '.html' uncachable:
acl HTML url_regex .html$ no_cache deny HTML
This example makes a specific URL uncachable:
acl XYZZY url_regex ^http://www.i.suck.com/foo.html$ no_cache deny XYZZY
This example caches nothing between the hours of 8AM to 11AM:
acl Morning time 08:00-11:00 no_cache deny Morning
In Squid-1.1, whether or not an object gets cached is controlled by the cache_stoplist, and cache_stoplist_pattern options. So, you may add:
cache_stoplist my.domain.comSpecifying uncachable objects by IP address is harder. The 1.1 patch page includes a patch called no-cache-local.patch which changes the behaviour of the local_ip and local_domain so that matching requests are NOT CACHED, in addition to being fetched directly.
Deleting an existing cache directory is not too difficult. Unfortunately, you can't simply change squid.conf and then reconfigure. You can't stop using a cache_dir while Squid is running. Also note that Squid requires at least one cache_dir to run.
/usr/local/squid/cachetmp ....If you add a new cache_dir you have to run squid -z to initialize that directory.
squid -k shutdown
squidclient mgr:storedir
The procedure is similar to recreate the directory.
% squid -z
by Dave J Woolley
If someone were to discover a buffer overrun bug in Squid and it runs as a user other than root, they can only corrupt the files writeable to that user, but if it runs a root, they can take over the whole machine. This applies to all programs that don't absolutely need root status, not just squid.
Here is a technique that was described by Radu Greab.
Start a second Squid server on an unused HTTP port (say 4128). This instance of Squid probably doesn't need a large disk cache. When this second server has finished reloading the disk store, swap the http_port values in the two squid.conf files. Set the original Squid to use port 5128, and the second one to use 3128. Next, run ``squid -k reconfigure'' for both Squids. New requests will go to the second Squid, now on port 3128 and the first Squid will finish handling its current requests. After a few minutes, it should be safe to fully shut down the first Squid and upgrade it. Later you can simply repeat this process in reverse.
Note: The information here is current for version 2.3.
Yes, you can specify multiple http_port lines in your squid.conf file. Squid attempts to bind() to each port that you specify. Sometimes Squid may not be able to bind to a port, either because of permissions or because the port is already in use. If Squid can bind to at least one port, then it will continue running. If it can not bind to any of the ports, then Squid stops.
With version 2.3 and later you can specify IP addresses and port numbers together (see the squid.conf comments).
Normally you cannot. Most TCP/IP stacks do not allow applications to create sockets with the local endpoint assigned to a foreign IP address. However, some folks have some patches to Linux that allow exactly that.
In this situation, you must ensure that all HTTP packets destined for the client IP addresses are routed to the Squid box. If the packets take another path, the real clients will send TCP resets to the origin servers, thereby breaking the connections.
Squid uses a lot of memory for performance reasons. It takes much, much longer to read something from disk than it does to read directly from memory.
A small amount of metadata for each cached object is kept in memory. This is the StoreEntry data structure. For Squid-2 this is 56-bytes on "small" pointer architectures (Intel, Sparc, MIPS, etc) and 88-bytes on "large" pointer architectures (Alpha). In addition, There is a 16-byte cache key (MD5 checksum) associated with each StoreEntry. This means there are 72 or 104 bytes of metadata in memory for every object in your cache. A cache with 1,000,000 objects therefore requires 72 MB of memory for metadata only. In practice it requires much more than that.
Squid-1.1 also uses a lot of memory to store in-transit objects. This version stores incoming objects only in memory, until the transfer is complete. At that point it decides whether or not to store the object on disk. This means that when users download large files, your memory usage will increase significantly. The squid.conf parameter maximum_object_size determines how much memory an in-transit object can consume before we mark it as uncachable. When an object is marked uncachable, there is no need to keep all of the object in memory, so the memory is freed for the part of the object which has already been written to the client. In other words, lowering maximum_object_size also lowers Squid-1.1 memory usage.
Other uses of memory by Squid include:
One way is to simply look at ps output on your system. For BSD-ish systems, you probably want to use the -u option and look at the VSZ and RSS fields:
wessels ~ 236% ps -axuhm
USER PID %CPU %MEM VSZ RSS TT STAT STARTED TIME COMMAND
squid 9631 4.6 26.4 141204 137852 ?? S 10:13PM 78:22.80 squid -NCYs
For SYSV-ish, you probably want to use the -l option.
When interpreting the ps output, be sure to check your ps
manual page. It may not be obvious if the reported numbers are kbytes,
or pages (usually 4 kb).
A nicer way to check the memory usage is with a program called top:
last pid: 20128; load averages: 0.06, 0.12, 0.11 14:10:58
46 processes: 1 running, 45 sleeping
CPU states: % user, % nice, % system, % interrupt, % idle
Mem: 187M Active, 1884K Inact, 45M Wired, 268M Cache, 8351K Buf, 1296K Free
Swap: 1024M Total, 256K Used, 1024M Free
PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND
9631 squid 2 0 138M 135M select 78:45 3.93% 3.93% squid
Finally, you can ask the Squid process to report its own memory usage. This is available on the Cache Manager info page. Your output may vary depending upon your operating system and Squid version, but it looks similar to this:
Resource usage for squid:
Maximum Resident Size: 137892 KB
Memory usage for squid via mstats():
Total space in arena: 140144 KB
Total free: 8153 KB 6%
If your RSS (Resident Set Size) value is much lower than your process size, then your cache performance is most likely suffering due to paging.
You might just have your cache_mem parameter set too high. See the `` What can I do to reduce Squid's memory usage?'' entry below.
When a process continually grows in size, without levelling off or slowing down, it often indicates a memory leak. A memory leak is when some chunk of memory is used, but not free'd when it is done being used.
Memory leaks are a real problem for programs (like Squid) which do all of their processing within a single process. Historically, Squid has had real memory leak problems. But as the software has matured, we believe almost all of Squid's memory leaks have been eliminated, and new ones are least easy to identify.
Memory leaks may also be present in your system's libraries, such as libc.a or even libmalloc.a. If you experience the ever-growing process size phenomenon, we suggest you first try an alternative malloc library.
The cache_mem parameter does NOT specify the maximum size of the process. It only specifies how much memory to use for caching ``hot'' (very popular) replies. Squid's actual memory usage is depends very strongly on your cache size (disk space) and your incoming request load. Reducing cache_mem will usually also reduce the process size, but not necessarily, and there are other ways to reduce Squid's memory usage (see below).
See also How much memory do I need in my Squid server?.
Note: This information is specific to Squid-1.1 versions
Look at your cachemgr.cgi Cache
Information page. For example:
Memory usage for squid via mallinfo():
Total space in arena: 94687 KB
Ordinary blocks: 32019 KB 210034 blks
Small blocks: 44364 KB 569500 blks
Holding blocks: 0 KB 5695 blks
Free Small blocks: 6650 KB
Free Ordinary blocks: 11652 KB
Total in use: 76384 KB 81%
Total free: 18302 KB 19%
Meta Data:
StoreEntry 246043 x 64 bytes = 15377 KB
IPCacheEntry 971 x 88 bytes = 83 KB
Hash link 2 x 24 bytes = 0 KB
URL strings = 11422 KB
Pool MemObject structures 514 x 144 bytes = 72 KB ( 70 free)
Pool for Request structur 516 x 4380 bytes = 2207 KB ( 2121 free)
Pool for in-memory object 6200 x 4096 bytes = 24800 KB ( 22888 free)
Pool for disk I/O 242 x 8192 bytes = 1936 KB ( 1888 free)
Miscellaneous = 2600 KB
total Accounted = 58499 KB
First note that mallinfo() reports 94M in ``arena.'' This
is pretty close to what top says (97M).
Of that 94M, 81% (76M) is actually being used at the moment. The
rest has been freed, or pre-allocated by malloc(3)
and not yet used.
Of the 76M in use, we can account for 58.5M (76%). There are some
calls to malloc(3) for which we can't account.
The Meta Data list gives the breakdown of where the
accounted memory has gone. 45% has gone to StoreEntry
and URL strings. Another 42% has gone to buffering hold objects
in VM while they are fetched and relayed to the clients (Pool
for in-memory object).
The pool sizes are specified by squid.conf parameters.
In version 1.0, these pools are somewhat broken: we keep a stack
of unused pages instead of freeing the block. In the Pool
for in-memory object, the unused stack size is 1/2 of
cache_mem. The Pool for disk I/O is
hardcoded at 200. For MemObject and Request
it's 1/8 of your system's FD_SETSIZE value.
If you need to lower your process size, we recommend lowering the
max object sizes in the 'http', 'ftp' and 'gopher' config lines.
You may also want to lower cache_mem to suit your
needs. But if you make cache_mem too low, then some
objects may not get saved to disk during high-load periods. Newer
Squid versions allow you to set memory_pools off to
disable the free memory pools.
We are not able to account for all memory that Squid uses. This would require excessive amounts of code to keep track of every last byte. We do our best to account for the major uses of memory.
Also, note that the malloc and free functions have their own overhead. Some additional memory is required to keep track of which chunks are in use, and which are free. Additionally, most operating systems do not allow processes to shrink in size. When a process gives up memory by calling free, the total process size does not shrink. So the process size really represents the maximum size your Squid process has reached.
Messages like "FATAL: xcalloc: Unable to allocate 4096 blocks of 1 bytes!" appear when Squid can't allocate more memory, and on most operating systems (inclusive BSD) there are only two possible reasons:
To tell if it is the second case, first rule out the first case and then monitor the size of the Squid process. If it dies at a certain size with plenty of swap left then the max data segment size is reached without no doubts.
The data segment size can be limited by two factors:
When squid starts it sets data and file ulimit's to the hard level. If you manually tune ulimit before starting Squid make sure that you set the hard limit and not only the soft limit (the default operation of ulimit is to only change the soft limit). root is allowed to raise the soft limit above the hard limit.
This command prints the hard limits:
ulimit -aH
This command sets the data size to unlimited:
ulimit -HSd unlimited
by Arjan de Vet
The default kernel limit on BSD/OS for datasize is 64MB (at least on 3.0 which I'm using).
Recompile a kernel with larger datasize settings:
maxusers 128 # Support for large inpcb hash tables, e.g. busy WEB servers. options INET_SERVER # support for large routing tables, e.g. gated with full Internet routing: options "KMEMSIZE=\(16*1024*1024\)" options "DFLDSIZ=\(128*1024*1024\)" options "DFLSSIZ=\(8*1024*1024\)" options "SOMAXCONN=128" options "MAXDSIZ=\(256*1024*1024\)"
See /usr/share/doc/bsdi/config.n for more info.
In /etc/login.conf I have this:
default:\
:path=/bin /usr/bin /usr/contrib/bin:\
:datasize-cur=256M:\
:openfiles-cur=1024:\
:openfiles-max=1024:\
:maxproc-cur=1024:\
:stacksize-cur=64M:\
:radius-challenge-styles=activ,crypto,skey,snk,token:\
:tc=auth-bsdi-defaults:\
:tc=auth-ftp-bsdi-defaults:
#
# Settings used by /etc/rc and root
# This must be set properly for daemons started as root by inetd as well.
# Be sure reset these values back to system defaults in the default class!
#
daemon:\
:path=/bin /usr/bin /sbin /usr/sbin:\
:widepasswords:\
:tc=default:
# :datasize-cur=128M:\
# :openfiles-cur=256:\
# :maxproc-cur=256:\
This should give enough space for a 256MB squid process.
by Duane Wessels
The procedure is almost identical to that for BSD/OS above. Increase the open filedescriptor limit in /sys/conf/param.c:
int maxfiles = 4096; int maxfilesperproc = 1024;Increase the maximum and default data segment size in your kernel config file, e.g. /sys/conf/i386/CONFIG:
options "MAXDSIZ=(512*1024*1024)" options "DFLDSIZ=(128*1024*1024)"We also found it necessary to increase the number of mbuf clusters:
options "NMBCLUSTERS=10240"And, if you have more than 256 MB of physical memory, you probably have to disable BOUNCE_BUFFERS (whatever that is), so comment out this line:
#options BOUNCE_BUFFERS #include support for DMA bounce buffers
Also, update limits in /etc/login.conf:
# Settings used by /etc/rc
#
daemon:\
:coredumpsize=infinity:\
:datasize=infinity:\
:maxproc=256:\
:maxproc-cur@:\
:memoryuse-cur=64M:\
:memorylocked-cur=64M:\
:openfiles=4096:\
:openfiles-cur@:\
:stacksize=64M:\
:tc=default:
And don't forget to run ``cap_mkdb /etc/login.conf'' after editing that file.
by Ong Beng Hui
To increase the data size for Digital UNIX, edit the file /etc/sysconfigtab
and add the entry...
proc:
per-proc-data-size=1073741824
Or, with csh, use the limit command, such as
> limit datasize 1024M
Editing /etc/sysconfigtab requires a reboot, but the limit command
doesn't.
When Squid is reconfigured (SIGHUP) or the logs are rotated (SIGUSR1), some of the helper processes (dnsserver) must be killed and restarted. If your system does not have enough virtual memory, the Squid process may not be able to fork to start the new helper processes. This is due to the UNIX way of starting child processes using the fork() system call which temporary duplicates the whole Squid process, and when rapidly starting many child processes such as on "squid -k rotate" the memory usage can temporarily grow to many times the normal memory usage due to several temporary copies of the whole process.
The best way to fix this is to increase your virtual memory by adding swap space. Normally your system uses raw disk partitions for swap space, but most operating systems also support swapping on regular files (Digital Unix excepted). See your system manual pages for swap, swapon, and mkfile. Alternatively you can use the sleep_after_fork directive to make Squid sleep a little while invoking helpers to allow the helper to start up before trying to start the next one. This can be helpful if you find that Squid sometimes fail to restart all helpers on "squid -k reconfigure".
If your cache performance is suffering because of memory limitations, you might consider buying more memory. But if that is not an option, There are a number of things to try:
Many users have found improved performance and memory utilization when linking Squid with an external malloc library. We recommend either GNU malloc, or dlmalloc.
To make Squid use GNU malloc follow these simple steps:
% gzip -dc malloc.tar.gz | tar xf - % cd malloc % vi Makefile # edit as needed % make
% su # cp malloc.a /usr/lib/libgnumalloc.a
# cp malloc.h /usr/include/gnumalloc.h
% make realclean % ./configure ... % make % make installNote, In later distributions, 'realclean' has been changed to 'distclean'. As the configure script runs, watch its output. You should find that it locates libgnumalloc.a and optionally gnumalloc.h.
dlmalloc has been written by Doug Lea. According to Doug:
This is not the fastest, most space-conserving, most portable, or most tunable malloc ever written. However it is among the fastest while also being among the most space-conserving, portable and tunable.
dlmalloc is included with the Squid-2 source distribution. To use this library, you simply give an option to the configure script:
% ./configure --enable-dlmalloc ...
As a rule of thumb on Squid uses approximately 10 MB of RAM per GB of the total of all cache_dirs (more on 64 bit servers such as Alpha), plus your cache_mem setting and about an additional 10-20MB. It is recommended to have at least twice this amount of physical RAM available on your Squid server. For a more detailed discussion on Squid's memory usage see the sections above.
The recommended extra RAM besides what is used by Squid is used by the operating system to improve disk I/O performance and by other applications or services running on the server. This will be true even of a server which runs Squid as the only tcp service, since there is a minimum level of memory needed for process management, logging, and other OS level routines.
If you have a low memory server, and a large disk, then you will not necessarily be able to use all the disk space, since as the cache fills the memory available will be insufficient, forcing Squid to swap out memory and affecting performance. A very large cache_dir total and insufficient physical RAM + Swap could cause Squid to stop functioning completely. The solution for larger caches is to get more physical RAM; allocating more to Squid via cache_mem will not help.
The cache manager (cachemgr.cgi) is a CGI utility for displaying statistics about the squid process as it runs. The cache manager is a convenient way to manage the cache and view statistics without logging into the server.
That depends on which web server you're using. Below you will find instructions for configuring the CERN and Apache servers to permit cachemgr.cgi usage.
EDITOR"S NOTE: readers are encouraged to submit instructions for configuration of cachemgr.cgi on other web server platforms, such as Netscape.
After you edit the server configuration files, you will probably
need to either restart your web server or or send it a SIGHUP signal
to tell it to re-read its configuration files.
When you're done configuring your web server, you'll connect to the cache manager with a web browser, using a URL such as:
http://www.example.com/Squid/cgi-bin/cachemgr.cgi/
First, you should ensure that only specified workstations can access the cache manager. That is done in your CERN httpd.conf, not in squid.conf.
Protection MGR-PROT {
Mask @(workstation.example.com)
}
Wildcards are acceptable, IP addresses are acceptable, and others can be added with a comma-separated list of IP addresses. There are many more ways of protection. Your server documentation has details.
You also need to add:
Protect /Squid/* MGR-PROT
Exec /Squid/cgi-bin/*.cgi /usr/local/squid/bin/*.cgi
This marks the script as executable to those in MGR-PROT.
First, make sure the cgi-bin directory you're using is listed with a
ScriptAlias in your Apache httpd.conf file like this:
ScriptAlias /Squid/cgi-bin/ /usr/local/squid/cgi-bin/It's probably a bad idea to
ScriptAlias
the entire usr/local/squid/bin/ directory where all the
Squid executables live.
Next, you should ensure that only specified workstations can access the cache manager. That is done in your Apache httpd.conf, not in squid.conf. At the bottom of httpd.conf file, insert:
<Location /Squid/cgi-bin/cachemgr.cgi>
order allow,deny
allow from workstation.example.com
</Location>
You can have more than one allow line, and you can allow domains or networks.
Alternately, cachemgr.cgi can be password-protected. You'd add the following to httpd.conf:
<Location /Squid/cgi-bin/cachemgr.cgi>
AuthUserFile /path/to/password/file
AuthGroupFile /dev/null
AuthName User/Password Required
AuthType Basic
require user cachemanager
</Location>
Consult the Apache documentation for information on using htpasswd to set a password for this ``user.''
by Francesco ``kinkie'' Chemolli
Notice: this is not how things would get best done with Roxen, but this what you need to do go adhere to the example. Also, knowledge of basic Roxen configuration is required.
This is what's required to start up a fresh Virtual Server, only serving the cache manager. If you already have some Virtual Server you wish to use to host the Cache Manager, just add a new CGI support module to it.
Create a new virtual server, and set it to host http://www.example.com/. Add to it at least the following modules:
In the CGI scripting support module, section Settings, change the following settings:
In section Security, set Patterns to:
allow ip=1.2.3.4where 1.2.3.4 is the IP address for workstation.example.com
Save the configuration, and you're done.
The default cache manager access configuration in squid.conf is:
acl manager proto cache_object
acl localhost src 127.0.0.1/255.255.255.255
acl all src 0.0.0.0/0.0.0.0
With the following rules:
http_access deny manager !localhost
http_access allow all
The first ACL is the most important as the cache manager program
interrogates squid using a special cache_object protocol.
Try it yourself by doing:
telnet mycache.example.com 3128 GET cache_object://mycache.example.com/info HTTP/1.0
The default ACLs say that if the request is for a
cache_object, and it isn't the local host, then deny
access; otherwise allow access.
In fact, only allowing localhost access means that on the
initial cachemgr.cgi form you can only specify the cache
host as localhost. We recommend the following:
acl manager proto cache_object
acl localhost src 127.0.0.1/255.255.255.255
acl example src 123.123.123.123/255.255.255.255
acl all src 0.0.0.0/0.0.0.0
Where 123.123.123.123 is the IP address of your web server.
Then modify the rules like this:
http_access allow manager localhost
http_access allow manager example
http_access deny manager
http_access allow all
If you're using miss_access, then don't forget to also add
a miss_access rule for the cache manager:
miss_access allow manager
The default ACLs assume that your web server is on the same machine
as squid. Remember that the connection from the cache
manager program to squid originates at the web server, not the
browser. So if your web server lives somewhere else, you should
make sure that IP address of the web server that has cachemgr.cgi
installed on it is in the example ACL above.
Always be sure to send a SIGHUP signal to squid
any time you change the squid.conf file.
If you ``drop'' the list box, and browse it, you will see that the password is only required to shutdown the cache, and the URL is required to refresh an object (i.e., retrieve it from its original source again) Otherwise these fields can be left blank: a password is not required to obtain access to the informational aspects of cachemgr.cgi.
See the cachemgr_passwd directive in squid.conf.
When you run configure use the --enable-cachemgr-hostname option:
% ./configure --enable-cachemgr-hostname=`hostname` ...
Note, if you do this after you already installed Squid before, you need to make sure cachemgr.cgi gets recompiled. For example:
% cd src % rm cachemgr.o cachemgr.cgi % make cachemgr.cgi
Then copy cachemgr.cgi to your HTTP server's cgi-bin directory.
Browsers and caches use TCP connections to retrieve web objects from web servers or caches. UDP connections are used when another cache using you as a sibling or parent wants to find out if you have an object in your cache that it's looking for. The UDP connections are ICP queries.
Don't worry. The default (and sensible) behavior of squid is to expire an object when it happens to overwrite it. It doesn't explicitly garbage collect (unless you tell it to in other ways).
Entry describing an object in the cache.
An entry in the DNS cache.
Link in the cache hash table structure.
The strings of the URLs themselves that map to an object number in the cache, allowing access to the StoreEntry.
Basically just like the log file in your cache directory:
PoolMemObject structuresPool for Request structuresPool for in-memory objectIf squid is much smaller than this field, run for cover! Something is very wrong, and you should probably restart squid.
Other?
Other is a default category to track objects which
don't fall into one of the defined categories.
Transfer KB/sec
column always zero?This column contains gross estimations of data transfer rates averaged over the entire time the cache has been running. These numbers are unreliable and mostly useless.
Object Count?
The number of objects of that type in the cache right now.
Max/Current/Min KB?
These refer to the size all the objects of this type have grown to/currently are/shrunk to.
I/O section about?
These are histograms on the number of bytes read from the network
per read(2) call. Somewhat useful for determining
maximum buffer sizes.
Objects section for?
Warning: this will download to your browser a list of every URL in the cache and statistics about it. It can be very, very large. Sometimes it will be larger than the amount of available memory in your client! You probably don't need this information anyway.
VM Objects section for?
VM Objects are the objects which are in Virtual Memory.
These are objects which are currently being retrieved and
those which were kept in memory for fast access (accelerator
mode).
AVG RTT mean?
Average Round Trip Time. This is how long on average after an ICP ping is sent that a reply is received.
A HIT means that the document was found in the cache. A MISS, that it wasn't found in the cache. A negative hit means that it was found in the cache, but it doesn't exist.
The hostname is the name that was requested to be resolved.
For the Flags column:
C Means positively cached.N Means negatively cached.P Means the request is pending being dispatched.D Means the request has been dispatched and we're waiting for an answer.L Means it is a locked entry because it represents a parent or sibling.The TTL column represents ``Time To Live'' (i.e., how long
the cache entry is valid). (May be negative if the document has
expired.)
The N column is the number of IP addresses from which
the cache has documents.
The rest of the line lists all the IP addresses that have been associated with that IP cache entry.
IPCache contains data for the Hostname to IP-Number mapping, and FQDNCache does it the other way round. For example:
IP Cache Contents:
Hostname Flags lstref TTL N [IP-Number]
gorn.cc.fh-lippe.de C 0 21581 1 193.16.112.73
lagrange.uni-paderborn.de C 6 21594 1 131.234.128.245
www.altavista.digital.com C 10 21299 4 204.123.2.75 ...
2/ftp.symantec.com DL 1583 -772855 0
Flags: C --> Cached
D --> Dispatched
N --> Negative Cached
L --> Locked
lstref: Time since last use
TTL: Time-To-Live until information expires
N: Count of addresses
FQDN Cache Contents:
IP-Number Flags TTL N Hostname
130.149.17.15 C -45570 1 andele.cs.tu-berlin.de
194.77.122.18 C -58133 1 komet.teuto.de
206.155.117.51 N -73747 0
Flags: C --> Cached
D --> Dispatched
N --> Negative Cached
L --> Locked
TTL: Time-To-Live until information expires
N: Count of names
This question was asked on the squid-users mailing list, to which there were three excellent replies.
You get a ``page fault'' when your OS tries to access something in memory which is actually swapped to disk. The term ``page fault'' while correct at the kernel and CPU level, is a bit deceptive to a user, as there's no actual error - this is a normal feature of operation.
Also, this doesn't necessarily mean your squid is swapping by that much. Most operating systems also implement paging for executables, so that only sections of the executable which are actually used are read from disk into memory. Also, whenever squid needs more memory, the fact that the memory was allocated will show up in the page faults.
However, if the number of faults is unusually high, and getting bigger, this could mean that squid is swapping. Another way to verify this is using a program called ``vmstat'' which is found on most UNIX platforms. If you run this as ``vmstat 5'' this will update a display every 5 seconds. This can tell you if the system as a whole is swapping a lot (see your local man page for vmstat for more information).
It is very bad for squid to swap, as every single request will be blocked until the requested data is swapped in. It is better to tweak the cache_mem and/or memory_pools setting in squid.conf, or switch to the NOVM versions of squid, than allow this to happen.
by Peter Wemm
There's two different operations at work, Paging and swapping. Paging is when individual pages are shuffled (either discarded or swapped to/from disk), while ``swapping'' generally means the entire process got sent to/from disk.
Needless to say, swapping a process is a pretty drastic event, and usually only reserved for when there's a memory crunch and paging out cannot free enough memory quickly enough. Also, there's some variation on how swapping is implemented in OS's. Some don't do it at all or do a hybrid of paging and swapping instead.
As you say, paging out doesn't necessarily involve disk IO, eg: text (code) pages are read-only and can simply be discarded if they are not used (and reloaded if/when needed). Data pages are also discarded if unmodified, and paged out if there's been any changes. Allocated memory (malloc) is always saved to disk since there's no executable file to recover the data from. mmap() memory is variable.. If it's backed from a file, it uses the same rules as the data segment of a file - ie: either discarded if unmodified or paged out.
There's also ``demand zeroing'' of pages as well that cause faults.. If you malloc memory and it calls brk()/sbrk() to allocate new pages, the chances are that you are allocated demand zero pages. Ie: the pages are not ``really'' attached to your process yet, but when you access them for the first time, the page fault causes the page to be connected to the process address space and zeroed - this saves unnecessary zeroing of pages that are allocated but never used.
The ``page faults with physical IO'' comes from the OS via getrusage(). It's highly OS dependent on what it means. Generally, it means that the process accessed a page that was not present in memory (for whatever reason) and there was disk access to fetch it. Many OS's load executables by demand paging as well, so the act of starting squid implicitly causes page faults with disk IO - however, many (but not all) OS's use ``read ahead'' and ``prefault'' heuristics to streamline the loading. Some OS's maintain ``intent queues'' so that pages can be selected as pageout candidates ahead of time. When (say) squid touches a freshly allocated demand zero page and one is needed, the OS can page out one of the candidates on the spot, causing a 'fault with physical IO' with demand zeroing of allocated memory which doesn't happen on many other OS's. (The other OS's generally put the process to sleep while the pageout daemon finds a page for it).
The meaning of ``swapping'' varies. On FreeBSD for example, swapping out is implemented as unlocking upages, kernel stack, PTD etc for aggressive pageout with the process. The only thing left of the process in memory is the 'struct proc'. The FreeBSD paging system is highly adaptive and can resort to paging in a way that is equivalent to the traditional swapping style operation (ie: entire process). FreeBSD also tries stealing pages from active processes in order to make space for disk cache. I suspect this is why setting 'memory_pools off' on the non-NOVM squids on FreeBSD is reported to work better - the VM/buffer system could be competing with squid to cache the same pages. It's a pity that squid cannot use mmap() to do file IO on the 4K chunks in it's memory pool (I can see that this is not a simple thing to do though, but that won't stop me wishing. :-).
by John Line
The comments so far have been about what paging/swapping figures mean in a ``traditional'' context, but it's worth bearing in mind that on some systems (Sun's Solaris 2, at least), the virtual memory and filesystem handling are unified and what a user process sees as reading or writing a file, the system simply sees as paging something in from disk or a page being updated so it needs to be paged out. (I suppose you could view it as similar to the operating system memory-mapping the files behind-the-scenes.)
The effect of this is that on Solaris 2, paging figures will also include file I/O. Or rather, the figures from vmstat certainly appear to include file I/O, and I presume (but can't quickly test) that figures such as those quoted by Squid will also include file I/O.
To confirm the above (which represents an impression from what I've read and observed, rather than 100% certain facts...), using an otherwise idle Sun Ultra 1 system system I just tried using cat (small, shouldn't need to page) to copy (a) one file to another, (b) a file to /dev/null, (c) /dev/zero to a file, and (d) /dev/zero to /dev/null (interrupting the last two with control-C after a while!), while watching with vmstat. 300-600 page-ins or page-outs per second when reading or writing a file (rather than a device), essentially zero in other cases (and when not cat-ing).
So ... beware assuming that all systems are similar and that paging figures represent *only* program code and data being shuffled to/from disk - they may also include the work in reading/writing all those files you were accessing...
You'll probably want to compare the number of page faults to the number of HTTP requests. If this ratio is close to, or exceeding 1, then Squid is paging too much.
This refers to ICP replies which Squid ignored, for one of these reasons:
Squid's access control scheme is relatively comprehensive and difficult for some people to understand. There are two different components: ACL elements, and access lists. An access list consists of an allow or deny action followed by a number of ACL elements.
Note: The information here is current for version 2.5.
Squid knows about the following types of ACL elements:
Notes:
Not all of the ACL elements can be used with all types of access lists (described below). For example, snmp_community is only meaningful when used with snmp_access. The src_as and dst_as types are only used in cache_peer_access access lists.
The arp ACL requires the special configure option --enable-arp-acl. Furthermore, the ARP ACL code is not portable to all operating systems. It works on Linux, Solaris, and some *BSD variants.
The SNMP ACL element and access list require the --enable-snmp configure option.
Some ACL elements can cause processing delays. For example, use of src_domain and srcdom_regex require a reverse DNS lookup on the client's IP address. This lookup adds some delay to the request.
Each ACL element is assigned a unique name. A named ACL element consists of a list of values. When checking for a match, the multiple values use OR logic. In other words, an ACL element is matched when any one of its values is a match.
You can't give the same name to two different types of ACL elements. It will generate a syntax error.
You can put different values for the same ACL name on different lines. Squid combines them into one list.
There are a number of different access lists:
Notes:
An access list rule consists of an allow or deny keyword, followed by a list of ACL element names.
An access list consists of one or more access list rules.
Access list rules are checked in the order they are written. List searching terminates as soon as one of the rules is a match.
If a rule has multiple ACL elements, it uses AND logic. In other words, all ACL elements of the rule must be a match in order for the rule to be a match. This means that it is possible to write a rule that can never be matched. For example, a port number can never be equal to both 80 AND 8000 at the same time.
To summarise the acl logics can be described as:
http_access allow|deny acl AND acl AND ...
OR
http_access allow|deny acl AND acl AND ...
OR
...
If none of the rules are matched, then the default action is the opposite of the last rule in the list. Its a good idea to be explicit with the default action. The best way is to thse the all ACL. For example:
acl all src 0/0 http_access deny all
Define an ACL that corresponds to your client's IP addresses. For example:
acl myclients src 172.16.5.0/24Next, allow those clients in the http_access list:
http_access allow myclients
acl someserver dstdomain .someserver.com no_cache deny someserver
As an example, we will assume that you would like to prevent users from accessing cooking recipes.
One way to implement this would be to deny access to any URLs that contain the words ``cooking'' or ``recipe.'' You would use these configuration lines:
acl Cooking1 url_regex cooking acl Recipe1 url_regex recipe http_access deny Cooking1 http_access deny Recipe1 http_access allow allThe url_regex means to search the entire URL for the regular expression you specify. Note that these regular expressions are case-sensitive, so a url containing ``Cooking'' would not be denied.
Another way is to deny access to specific servers which are known to hold recipes. For example:
acl Cooking2 dstdomain www.gourmet-chef.com
http_access deny Cooking2
http_access allow all
The dstdomain means to search the hostname in the URL for the
string ``www.gourmet-chef.com.''
Note that when IP addresses are used in URLs (instead of domain names),
Squid-1.1 implements relaxed access controls. If the a domain name
for the IP address has been saved in Squid's ``FQDN cache,'' then
Squid can compare the destination domain against the access controls.
However, if the domain is not immediately available, Squid allows
the request and makes a lookup for the IP address so that it may
be available for future reqeusts.
You can use ident lookups to allow specific users access to your cache. This requires that an ident server process runs on the user's machine(s). In your squid.conf configuration file you would write something like this:
ident_lookup_access allow all acl friends ident kim lisa frank joe http_access allow friends http_access deny all
Another option is to use proxy-authentication. In this scheme, you assign usernames and passwords to individuals. When they first use the proxy they are asked to authenticate themselves by entering their username and password.
In Squid v2 this authentication is hanled via external processes. For information on how to configure this, please see Configuring Proxy Authentication.
Pedro L Orso has adapted the Apache's htpasswd into a CGI program called chpasswd.cgi.
You can use the ident_access directive to control for which hosts Squid will issue ident lookup requests.
Additionally, if you use a ident ACL in squid conf, then Squid will make sure an ident lookup is performed while evaluating the acl even if iden_access does not indicate ident lookups should be performed.
However, Squid does not wait for the lookup to complete unless the ACL rules require it. Consider this configuration:
acl host1 src 10.0.0.1 acl host2 src 10.0.0.2 acl pals ident kim lisa frank joe http_access allow host1 http_access allow host2 palsRequests coming from 10.0.0.1 will be allowed immediately because there are no user requirements for that host. However, requests from 10.0.0.2 will be allowed only after the ident lookup completes, and if the username is in the set kim, lisa, frank, or joe.
You've probably noticed (and been frustrated by) the fact that you cannot combine access controls with terms like ``and'' or ``or.'' These operations are already built in to the access control scheme in a fundamental way which you must understand.
For example, the following access control configuration will never work:
acl ME src 10.0.0.1 acl YOU src 10.0.0.2 http_access allow ME YOUIn order for the request to be allowed, it must match the ``ME'' acl AND the ``YOU'' acl. This is impossible because any IP address could only match one or the other. This should instead be rewritten as:
acl ME src 10.0.0.1 acl YOU src 10.0.0.2 http_access allow ME http_access allow YOUOr, alternatively, this would also work:
acl US src 10.0.0.1 10.0.0.2 http_access allow US
I have read through my squid.conf numerous times, spoken to my neighbors, read the FAQ and Squid Docs and cannot for the life of me work out why the following will not work.
I can successfully access cachemgr.cgi from our web server machine here, but I would like to use MRTG to monitor various aspects of our proxy. When I try to use 'squidclient' or GET cache_object from the machine the proxy is running on, I always get access denied.
acl manager proto cache_object acl localhost src 127.0.0.1/255.255.255.255 acl server src 1.2.3.4/255.255.255.255 acl all src 0.0.0.0/0.0.0.0 acl ourhosts src 1.2.0.0/255.255.0.0 http_access deny manager !localhost !server http_access allow ourhosts http_access deny all
The intent here is to allow cache manager requests from the localhost and server addresses, and deny all others. This policy has been expressed here:
http_access deny manager !localhost !server
The problem here is that for allowable requests, this access rule is not matched. For example, if the source IP address is localhost, then ``!localhost'' is false and the access rule is not matched, so Squid continues checking the other rules. Cache manager requests from the server address work because server is a subset of ourhosts and the second access rule will match and allow the request. Also note that this means any cache manager request from ourhosts would be allowed.
To implement the desired policy correctly, the access rules should be rewritten as
http_access allow manager localhost http_access allow manager server http_access deny manager http_access allow ourhosts http_access deny allIf you're using miss_access, then don't forget to also add a miss_access rule for the cache manager:
miss_access allow manager
You may be concerned that the having five access rules instead of three may have an impact on the cache performance. In our experience this is not the case. Squid is able to handle a moderate amount of access control checking without degrading overall performance. You may like to verify that for yourself, however.
For the srcdomain ACL type, Squid does a reverse lookup of the client's IP address and checks the result with the domains given on the acl line. With the src ACL type, Squid converts hostnames to IP addresses at startup and then only compares the client's IP address. The src ACL is preferred over srcdomain because it does not require address-to-name lookups for each request.
If ACLs are giving you problems and you don't know why they aren't working, you can use this tip to debug them.
In squid.conf enable debugging for section 33 at level 2. For example:
debug_options ALL,1 33,2Then restart or reconfigure squid.
From now on, your cache.log should contain a line for every request that explains if it was allowed, or denied, and which ACL was the last one that it matched.
If this does not give you sufficient information to nail down the problem you can also enable detailed debug information on ACL processing
debug_options ALL,1 33,2 28,9Then restart or reconfigure squid as above.
From now on, your cache.log should contain detailed traces of all access list processing. Be warned that this can be quite some lines per request.
See also 11.20 Debug Squid
The problem...
[ Parents ] / \ / \ [ Proxy A ] --- [ Proxy B ] | | USERProxy A sends and ICP query to Proxy B about an object, Proxy B replies with an ICP_HIT. Proxy A forwards the HTTP request to Proxy B, but does not pass on the authentication details, therefore the HTTP GET from Proxy A fails.
Only ONE proxy cache in a chain is allowed to ``use'' the Proxy-Authentication request header. Once the header is used, it must not be passed on to other proxies.
Therefore, you must allow the neighbor caches to request from each other without proxy authentication. This is simply accomplished by listing the neighbor ACL's first in the list of http_access lines. For example:
acl proxy-A src 10.0.0.1 acl proxy-B src 10.0.0.2 acl user_passwords proxy_auth /tmp/user_passwds http_access allow proxy-A http_access allow proxy-B http_access allow user_passwords http_access deny all
acl GOOD dst 10.0.0.1 acl BAD dst 0.0.0.0/0.0.0.0 http_access allow GOOD http_access deny BAD
If you are using Squid-2.4 or later then keep in mind that dstdomain acls uses different syntax for exact host matches and entire domain matches. www.example.com matches the exact host www.example.com, while .example.com matches the entire domain example.com (including example.com alone)
There is also subtle issues if your dstdomain ACLs contains matches for both an exact host in a domain and the whole domain where both are in the same domain (i.e. both www.example.com and .example.com). Depending on how your data is ordered this may cause only the most specific of these (e.g. www.example.com) to be used.
NOTE: Current Squid versions (as of Squid-2.4) will warn you when this kind of configuration is used. If your Squid does not warn you while reading the configuration file you do not have the problem described below. Also the configuration here uses the dstdomain syntax of Squid-2.1 or earlier.. (2.2 and later needs to have domains prefixed by a dot)
There is a subtle problem with domain-name based access controls when a single ACL element has an entry that is a subdomain of another entry. For example, consider this list:
acl FOO dstdomain boulder.co.us vail.co.us co.us
In the first place, the above list is simply wrong because the first two (boulder.co.us and vail.co.us) are unnecessary. Any domain name that matches one of the first two will also match the last one (co.us). Ok, but why does this happen?
The problem stems from the data structure used to index domain names in an access control list. Squid uses Splay trees for lists of domain names. As other tree-based data structures, the searching algorithm requires a comparison function that returns -1, 0, or +1 for any pair of keys (domain names). This is similar to the way that strcmp() works.
The problem is that it is wrong to say that co.us is greater-than, equal-to, or less-than boulder.co.us.
For example, if you said that co.us is LESS than fff.co.us, then the Splay tree searching algorithm might never discover co.us as a match for kkk.co.us.
similarly, if you said that co.us is GREATER than fff.co.us, then the Splay tree searching algorithm might never discover co.us as a match for bbb.co.us.
The bottom line is that you can't have one entry that is a subdomain of another. Squid-2.2 will warn you if it detects this condition.
It is dangerous to allow Squid to connect to certain port numbers. For example, it has been demonstrated that someone can use Squid as an SMTP (email) relay. As I'm sure you know, SMTP relays are one of the ways that spammers are able to flood our mailboxes. To prevent mail relaying, Squid denies requests when the URL port number is 25. Other ports should be blocked as well, as a precaution.
There are two ways to filter by port number: either allow specific ports, or deny specific ports. By default, Squid does the first. This is the ACL entry that comes in the default squid.conf:
acl Safe_ports port 80 21 443 563 70 210 1025-65535 http_access deny !Safe_portsThe above configuration denies requests when the URL port number is not in the list. The list allows connections to the standard ports for HTTP, FTP, Gopher, SSL, WAIS, and all non-priveleged ports.
Another approach is to deny dangerous ports. The dangerous port list should look something like:
acl Dangerous_ports 7 9 19 22 23 25 53 109 110 119 http_access deny Dangerous_ports...and probably many others.
Please consult the /etc/services file on your system for a list of known ports and protocols.
Note: The information here is current for version 2.2.
No, it does not.
This example allows only the special_client to access the special_url. Any other client that tries to access the special_url is denied.
acl special_client src 10.1.2.3 acl special_url url_regex ^http://www.squid-cache.org/Doc/FAQ/$ http_access allow special_client special_url http_access deny special_url
Let's say you have two workstations that should only be allowed access to the Internet during working hours (8:30 - 17:30). You can use something like this:
acl FOO src 10.1.2.3 10.1.2.4 acl WORKING time MTWHF 08:30-17:30 http_access allow FOO WORKING http_access deny FOO
acl USER1 proxy_auth Dick acl USER2 proxy_auth Jane acl DAY time 06:00-18:00 http_access allow USER1 DAY http_access deny USER1 http_access allow USER2 !DAY http_access deny USER2
Note: The information here is current for version 2.3.
The following ACL entry gives inconsistent or unexpected results:
acl restricted src 10.0.0.128/255.0.0.128 10.85.0.0/16The reason is that IP access lists are stored in ``splay'' tree data structures. These trees require the keys to be sortable. When you use a complicated, or non-standard, netmask (255.0.0.128), it confuses the function that compares two address/mask pairs.
The best way to fix this problem is to use separate ACL names for each ACL value. For example, change the above to:
acl restricted1 src 10.0.0.128/255.0.0.128 acl restricted2 src 10.85.0.0/16
Then, of course, you'll have to rewrite your http_access lines as well.
Yes, for some operating systes. Squid calls these ``ARP ACLs'' and they are supported on Linux, Solaris, and probably BSD variants.
NOTE: Squid can only determine the MAC address for clients that are on the same subnet. If the client is on a different subnet, then Squid can not find out its MAC address.
To use ARP (MAC) access controls, you first need to compile in the optional code. Do this with the --enable-arp-acl configure option:
% ./configure --enable-arp-acl ... % make clean % makeIf src/acl.c doesn't compile, then ARP ACLs are probably not supported on your system.
If everything compiles, then you can add some ARP ACL lines to your squid.conf:
acl M1 arp 01:02:03:04:05:06 acl M2 arp 11:12:13:14:15:16 http_access allow M1 http_access allow M2 http_access deny all
See 1.9 I set up my access controls, but they don't work! why? and 11.20 Debugging Squid.
Yes, use the maxconn ACL type in conjunction with http_access deny. For example:
acl losers src 1.2.3.0/24 acl 5CONN maxconn 5 http_access deny 5CONN losers
Given the above configuration, when a client whose source IP address is in the 1.2.3.0/24 subnet tries to establish 6 or more connections at once, Squid returns an error page. Unless you use the deny_info feature, the error message will just say ``access denied.''
The maxconn ACL requires the client_db feature. If you've disabled client_db (for example with client_db off) then maxconn ALCs will not work.
Note, the maxconn ACL type is kind of tricky because it uses less-than comparison. The ACL is a match when the number of established connections is greater than the value you specify. Because of that, you don't want to use the maxconn ACL with http_access allow.
Also note that you could use maxconn in conjunction with a user type (ident, proxy_auth), rather than an IP address type.
In Squid-2.3 we changed the way that Squid matches subdomains. There is a difference between .foo.com and foo.com. The first matches any domain in foo.com, while the latter matches only ``foo.com'' exactly. So if you want to deny bar.foo.com, you should write
acl yuck dstdomain .foo.com http_access deny yuck
You can customize the existing error messages as described in Customizable Error Messages. You can also create new error messages and use these in conjunction with the deny_info option.
For example, lets say you want your users to see a special message when they request something that matches your pornography list. First, create a file named ERR_NO_PORNO in the /usr/local/squid/etc/errors directory. That file might contain something like this:
<p> Our company policy is to deny requests to known porno sites. If you feel you've received this message in error, please contact the support staff (support@this.company.com, 555-1234).
Next, set up your access controls as follows:
acl porn url_regex "/usr/local/squid/etc/porno.txt" deny_info ERR_NO_PORNO porn http_access deny porn (additional http_access lines ...)
Squid by defaults uses GMT as timestamp in all geenrated error messages. This to allow the cache to participate in a hierarchy of caches in different timezones without risking confusion about what the time is.
To change the timestamp in Squid generated error messages you must change the Squid signature. See Customizable Error Messages. The signature by defaults uses %T as timestamp, but if you like then you can use %t instead for a timestamp using local time zone.
by Adam Aube
Squid can read ACL parameters from an external file. To do this, first place the acl parameters, one per line, in a file. Then, on the ACL line in squid.conf, put the full path to the file in double quotes.
For example, instead of:
acl trusted_users proxy_auth john jane jim
you would have:
acl trusted_users proxy_auth "/usr/local/squid/etc/trusted_users.txt"
Inside trusted_users.txt, there is:
john jane jim
You may need to set up the http_access option to allow requests from your IP addresses. Please see the Access Controls section for information about that.
If squid is in httpd-accelerator mode, it will accept normal HTTP requests and forward them to a HTTP server, but it will not honor proxy requests. If you want your cache to also accept proxy-HTTP requests then you must enable this feature:
httpd_accel_with_proxy onAlternately, you may have misconfigured one of your ACLs. Check the access.log and squid.conf files for clues.
local_domain to work; Squid is caching the objects from my local servers.
The local_domain directive does not prevent local
objects from being cached. It prevents the use of sibling caches
when fetching local objects. If you want to prevent objects from
being cached, use the cache_stoplist or http_stop
configuration options (depending on your version).
Connection Refused when the cache tries to retrieve an object located on a sibling, even though the sibling thinks it delivered the object to my cache.
If the HTTP port number is wrong but the ICP port is correct you
will send ICP queries correctly and the ICP replies will fool your
cache into thinking the configuration is correct but large objects
will fail since you don't have the correct HTTP port for the sibling
in your squid.conf file. If your sibling changed their
http_port, you could have this problem for some time
before noticing.
If you see the Too many open files error message, you
are most likely running out of file descriptors. This may be due
to running Squid on an operating system with a low filedescriptor
limit. This limit is often configurable in the kernel or with
other system tuning tools. There are two ways to run out of file
descriptors: first, you can hit the per-process limit on file
descriptors. Second, you can hit the system limit on total file
descriptors for all processes.
Linux kernel 2.2.12 and later supports "unlimited" number of open files without patching. So does most of glibc-2.1.1 and later (all areas touched by Squid is safe from what I can tell, even more so in later glibc releases). But you still need to take some actions as the kernel defaults to only allow processes to use up to 1024 filedescriptors, and Squid picks up the limit at build time.
Alternatively you can
If running things as root is not an option then get your sysadmin to install a the needed ulimit command in /etc/inittscript (see man initscript), install a patched kernel where INR_OPEN in include/linux/fs.h is changed to at least the amount you need or have them install a small suid program which sets the limit (see link below).
More information can be found from Henriks How to get many filedescriptors on Linux 2.2.X and later page.
Add the following to your /etc/system file and reboot to increase your maximum file descriptors per process:
set rlim_fd_max = 4096
Next you should re-run the configure script
in the top directory so that it finds the new value.
If it does not find the new limit, then you might try
editing include/autoconf.h and setting
#define DEFAULT_FD_SETSIZE by hand. Note that
include/autoconf.h is created from autoconf.h.in
every time you run configure. Thus, if you edit it by
hand, you might lose your changes later on.
Jens-S. Voeckler advises that you should NOT change the default soft limit (rlim_fd_cur) to anything larger than 256. It will break other programs, such as the license manager needed for the SUN workshop compiler. Jens-S. also says that it should be safe to raise the limit for the Squid process as high as 16,384 except that there may be problems duruing reconfigure or logrotate if all of the lower 256 filedescriptors are in use at the time or rotate/reconfigure.
Do sysctl -a and look for the value of
kern.maxfilesperproc.
sysctl -w kern.maxfiles=XXXX sysctl -w kern.maxfilesperproc=XXXXWarning: You probably want
maxfiles
> maxfilesperproc if you're going to be pushing the
limit.I don't think there is a formal upper limit inside the kernel. All the data structures are dynamically allocated. In practice there might be unintended metaphenomena (kernel spending too much time searching tables, for example).
For most BSD-derived systems (SunOS, 4.4BSD, OpenBSD, FreeBSD, NetBSD, BSD/OS, 386BSD, Ultrix) you can also use the ``brute force'' method to increase these values in the kernel (requires a kernel rebuild):
Do pstat -T and look for the files
value, typically expressed as the ratio of currentmaximum/.
One way is to increase the value of the maxusers variable
in the kernel configuration file and build a new kernel. This method
is quick and easy but also has the effect of increasing a wide variety of
other variables that you may not need or want increased.
Another way is to find the param.c file in your kernel
build area and change the arithmetic behind the relationship between
maxusers and the maximum number of open files.
Change the value of nfile in usr/kvm/sys/conf.common/param.c/tt> by altering this equation:
int nfile = 16 * (NPROC + 16 + MAXUSERS) / 10 + 64;Where
NPROC is defined by:
#define NPROC (10 + 16 * MAXUSERS)
Very similar to SunOS, edit /usr/src/sys/conf/param.c
and alter the relationship between maxusers and the
maxfiles and maxfilesperproc variables:
int maxfiles = NPROC*2; int maxfilesperproc = NPROC*2;Where
NPROC is defined by:
#define NPROC (20 + 16 * MAXUSERS)
The per-process limit can also be adjusted directly in the kernel
configuration file with the following directive:
options OPEN_MAX=128
Edit /usr/src/sys/conf/param.c and adjust the
maxfiles math here:
int maxfiles = 3 * (NPROC + MAXUSERS) + 80;Where
NPROC is defined by:
#define NPROC (20 + 16 * MAXUSERS)
You should also set the OPEN_MAX value in your kernel
configuration file to change the per-process limit.
NOTE: After you rebuild/reconfigure your kernel with more filedescriptors, you must then recompile Squid. Squid's configure script determines how many filedescriptors are available, so you must make sure the configure script runs again as well. For example:
cd squid-1.1.x
make realclean
./configure --prefix=/usr/local/squid
make
For example:
97/01/23 22:31:10| Removed 1 of 9 objects from bucket 3913 97/01/23 22:33:10| Removed 1 of 5 objects from bucket 4315 97/01/23 22:35:40| Removed 1 of 14 objects from bucket 6391
These log entries are normal, and do not indicate that squid has
reached cache_swap_high.
Consult your cache information page in cachemgr.cgi for a line like this:
Storage LRU Expiration Age: 364.01 days
Objects which have not been used for that amount of time are removed as
a part of the regular maintenance. You can set an upper limit on the
LRU Expiration Age value with reference_age in the config
file.
Why, yes you can! Select the following menus:
This will bring up a box with icons for your various services. One of them should be a little ftp ``folder.'' Double click on this.
You will then have to select the server (there should only be one) Select that and then choose ``Properties'' from the menu and choose the ``directories'' tab along the top.
There will be an option at the bottom saying ``Directory listing style.'' Choose the ``Unix'' type, not the ``MS-DOS'' type.
--Oskar Pearson <oskar@is.co.za>
You are receiving ICP MISSes (via UDP) from a parent or sibling cache whose IP address your cache does not know about. This may happen in two situations.
on your parent squid.conf:
udp_outgoing_address proxy.parent.comon your squid.conf:
cache_peer proxy.parent.com parent 3128 3130
The standards for naming hosts ( RFC 952, RFC 1101) do not allow underscores in domain names:
A "name" (Net, Host, Gateway, or Domain name) is a text string up to 24 characters drawn from the alphabet (A-Z), digits (0-9), minus sign (-), and period (.).The resolver library that ships with recent versions of BIND enforces this restriction, returning an error for any host with underscore in the hostname. The best solution is to complain to the hostmaster of the offending site, and ask them to rename their host.
See also the comp.protocols.tcp-ip.domains FAQ.
Some people have noticed that RFC 1033 implies that underscores are allowed. However, this is an informational RFC with a poorly chosen example, and not a standard by any means.
See the above question. The underscore character is not valid for hostnames.
Some DNS resolvers allow the underscore, so yes, the hostname might work fine when you don't use Squid.
To make Squid allow underscores in hostnames, re-run the configure script with this option:
% ./configure --enable-underscores ...and then recompile:
% make clean % make
The answer to this is somewhat complicated, so please hold on. NOTE: most of this text is taken from ICP and the Squid Web Cache.
An ICP query does not include any parent or sibling designation,
so the receiver really has no indication of how the peer
cache is configured to use it. This issue becomes important
when a cache is willing to serve cache hits to anyone, but only
handle cache misses for its paying users or customers. In other
words, whether or not to allow the request depends on if the
result is a hit or a miss. To accomplish this,
Squid acquired the miss_access feature
in October of 1996.
The necessity of ``miss access'' makes life a little bit complicated,
and not only because it was awkward to implement. Miss access
means that the ICP query reply must be an extremely accurate prediction
of the result of a subsequent HTTP request. Ascertaining
this result is actually very hard, if not impossible to
do, since the ICP request cannot convey the
full HTTP request.
Additionally, there are more types of HTTP request results than there
are for ICP. The ICP query reply will either be a hit or miss.
However, the HTTP request might result in a ``304 Not Modified'' reply
sent from the origin server. Such a reply is not strictly a hit since the peer
needed to forward a conditional request to the source. At the same time,
its not strictly a miss either since the local object data is still valid,
and the Not-Modified reply is quite small.
One serious problem for cache hierarchies is mismatched freshness parameters. Consider a cache C using ``strict'' freshness parameters so its users get maximally current data. C has a sibling S with less strict freshness parameters. When an object is requested at C, C might find that S already has the object via an ICP query and ICP HIT response. C then retrieves the object from S.
In an HTTP/1.0 world, C (and C's client) will receive an object that was never subject to its local freshness rules. Neither HTTP/1.0 nor ICP provides any way to ask only for objects less than a certain age. If the retrieved object is stale by Cs rules, it will be removed from Cs cache, but it will subsequently be fetched from S so long as it remains fresh there. This configuration miscoupling problem is a significant deterrent to establishing both parent and sibling relationships.
HTTP/1.1 provides numerous request headers to specify freshness
requirements, which actually introduces
a different problem for cache hierarchies: ICP
still does not include any age information, neither in query nor
reply. So S may return an ICP HIT if its
copy of the object is fresh by its configuration
parameters, but the subsequent HTTP request may result
in a cache miss due to any
Cache-control: headers originated by C or by
C's client. Situations now emerge where the ICP reply
no longer matches the HTTP request result.
In the end, the fundamental problem is that the ICP query does not provide enough information to accurately predict whether the HTTP request will be a hit or miss. In fact, the current ICP Internet Draft is very vague on this subject. What does ICP HIT really mean? Does it mean ``I know a little about that URL and have some copy of the object?'' Or does it mean ``I have a valid copy of that object and you are allowed to get it from me?''
So, what can be done about this problem? We really need to change ICP so that freshness parameters are included. Until that happens, the members of a cache hierarchy have only two options to totally eliminate the ``access denied'' messages from sibling caches:
refresh_rules parameters.miss_access at all. Promise your sibling cache
administrator that your cache is properly configured and that you
will not abuse their generosity. The sibling cache administrator can
check his log files to make sure you are keeping your word.This means that another processes is already listening on port 8080 (or whatever you're using). It could mean that you have a Squid process already running, or it could be from another program. To verify, use the netstat command:
netstat -naf inet | grep LISTENThat will show all sockets in the LISTEN state. You might also try
netstat -naf inet | grep 8080If you find that some process has bound to your port, but you're not sure which process it is, you might be able to use the excellent lsof program. It will show you which processes own every open file descriptor on your system.
This means that the client socket was closed by the client
before Squid was finished sending data to it. Squid detects this
by trying to read(2) some data from the socket. If the
read(2) call fails, then Squid konws the socket has been
closed. Normally the read(2) call returns ECONNRESET: Connection reset by peer
and these are NOT logged. Any other error messages (such as
EPIPE: Broken pipe are logged to cache.log. See the ``intro'' of
section 2 of your Unix manual for a list of all error codes.
These are caused by misbehaving Web clients attempting to use persistent connections. Squid-1.1 does not support persistent connections.
Version 2.5 will support Microsoft NTLM authentication. However, there are some limits on our support: We cannot proxy connections to a origin server that use NTLM authentication, but we can act as a web accelerator or proxy server and authenticate the client connection using NTLM.
We support NT4, Samba, and Windows 2000 Domain Controllers. For more information see winbind .
Why we cannot proxy NTLM even though we can use it. Quoting from summary at the end of the browser authentication section in this article:
In summary, Basic authentication does not require an implicit end-to-end state, and can therefore be used through a proxy server. Windows NT Challenge/Response authentication requires implicit end-to-end state and will not work through a proxy server.
Squid transparently passes the NTLM request and response headers between clients and servers. NTLM relies on a single end-end connection (possibly with men-in-the-middle, but a single connection every step of the way. This implies that for NTLM authentication to work at all with proxy caches, the proxy would need to tightly link the client-proxy and proxy-server links, as well as understand the state of the link at any one time. NTLM through a CONNECT might work, but we as far as we know that hasn't been implemented by anyone, and it would prevent the pages being cached - removing the value of the proxy.
NTLM authentication is carried entirely inside the HTTP protocol, but is not a true HTTP authentication protocol and is different from Basic and Digest authentication in many ways.
The reasons why it is not implemented in Netscape is probably:
This message was received at squid-bugs:
If you have only one parent, configured as:cache_peer xxxx parent 3128 3130 no-query defaultnothing is sent to the parent; neither UDP packets, nor TCP connections.
Simply adding default to a parent does not force all requests to be sent to that parent. The term default is perhaps a poor choice of words. A default parent is only used as a last resort. If the cache is able to make direct connections, direct will be preferred over default. If you want to force all requests to your parent cache(s), use the never_direct option:
acl all src 0.0.0.0/0.0.0.0 never_direct allow all
``Hot Mail'' is proxy-unfriendly and requires all requests to come from the same IP address. You can fix this by adding to your squid.conf:
hierarchy_stoplist hotmail.com
This is most likely because Squid is using more memory than it should be for your system. When the Squid process becomes large, it experiences a lot of paging. This will very rapidly degrade the performance of Squid. Memory usage is a complicated problem. There are a number of things to consider.
Then, examine the Cache Manager Info ouput and look at these two lines:
Number of HTTP requests received: 121104 Page faults with physical i/o: 16720Note, if your system does not have the getrusage() function, then you will not see the page faults line.
Divide the number of page faults by the number of connections. In this case 16720/121104 = 0.14. Ideally this ratio should be in the 0.0 - 0.1 range. It may be acceptable to be in the 0.1 - 0.2 range. Above that, however, and you will most likely find that Squid's performance is unacceptably slow.
If the ratio is too high, you will need to make some changes to lower the amount of memory Squid uses.
See also How much memory do I need in my Squid server?.
This could be a permission problem. Does the Squid userid have permission to execute the dnsserver program?
You might also try testing dnsserver from the command line:
> echo oceana.nlanr.net | ./dnsserverShould produce something like:
$name oceana.nlanr.net $h_name oceana.nlanr.net $h_len 4 $ipcount 1 132.249.40.200 $aliascount 0 $ttl 82067 $end
Bug reports for Squid should be registered in our bug database. Any bug report must include
Please note that bug reports are only processed if they can be reproduced or identified in the current STABLE or development versions of Squid. If you are running an older version of Squid the first response will be to ask you to upgrade unless the developer who looks at your bug report immediately can identify that the bug also exists in the current versions. It should also be noted that any patches provided by the Squid developer team will be to the current STABLE version even if you run an older version.
There are two conditions under which squid will exit abnormally and generate a coredump. First, a SIGSEGV or SIGBUS signal will cause Squid to exit and dump core. Second, many functions include consistency checks. If one of those checks fail, Squid calls abort() to generate a core dump.
Many people report that Squid doesn't leave a coredump anywhere. This may be due to one of the following reasons:
# sysctl -w kern.sugid_coredump=1
Resource Limits: These limits can usually be changed in shell scripts. The command to change the resource limits is usually either limit or limits. Sometimes it is a shell-builtin function, and sometimes it is a regular program. Also note that you can set resource limits in the /etc/login.conf file on FreeBSD and maybe other BSD systems.
To change the coredumpsize limit you might use a command like:
limit coredumpsize unlimitedor
limits coredump unlimited
Debugging Symbols: To see if your Squid binary has debugging symbols, use this command:
% nm /usr/local/squid/bin/squid | headThe binary has debugging symbols if you see gobbledegook like this:
0812abec B AS_tree_head 080a7540 D AclMatchedName 080a73fc D ActionTable 080908a4 r B_BYTES_STR 080908bc r B_GBYTES_STR 080908ac r B_KBYTES_STR 080908b4 r B_MBYTES_STR 080a7550 D Biggest_FD 08097c0c R CacheDigestHashFuncCount 08098f00 r CcAttrsThere are no debugging symbols if you see this instead:
/usr/local/squid/bin/squid: no symbolsDebugging symbols may have been removed by your install program. If you look at the squid binary from the source directory, then it might have the debugging symbols.
Coredump Location: The core dump file will be left in one of the following locations:
2000/03/14 00:12:36| Set Current Directory to /usr/local/squid/cacheIf you cannot find a core file, then either Squid does not have permission to write in its current directory, or perhaps your shell limits are preventing the core file from being written.
Often you can get a coredump if you run Squid from the command line like this (csh shells and clones):
% limit core un % /usr/local/squid/bin/squid -NCd1
Once you have located the core dump file, use a debugger such as dbx or gdb to generate a stack trace:
tirana-wessels squid/src 270% gdb squid /T2/Cache/core GDB is free software and you are welcome to distribute copies of it under certain conditions; type "show copying" to see the conditions. There is absolutely no warranty for GDB; type "show warranty" for details. GDB 4.15.1 (hppa1.0-hp-hpux10.10), Copyright 1995 Free Software Foundation, Inc... Core was generated by `squid'. Program terminated with signal 6, Aborted. [...] (gdb) where #0 0xc01277a8 in _kill () #1 0xc00b2944 in _raise () #2 0xc007bb08 in abort () #3 0x53f5c in __eprintf (string=0x7b037048 "", expression=0x5f <Address 0x5f out of bounds>, line=8, filename=0x6b <Address 0x6b out of bounds>) #4 0x29828 in fd_open (fd=10918, type=3221514150, desc=0x95e4 "HTTP Request") at fd.c:71 #5 0x24f40 in comm_accept (fd=2063838200, peer=0x7b0390b0, me=0x6b) at comm.c:574 #6 0x23874 in httpAccept (sock=33, notused=0xc00467a6) at client_side.c:1691 #7 0x25510 in comm_select_incoming () at comm.c:784 #8 0x25954 in comm_select (sec=29) at comm.c:1052 #9 0x3b04c in main (argc=1073745368, argv=0x40000dd8) at main.c:671
If possible, you might keep the coredump file around for a day or two. It is often helpful if we can ask you to send additional debugger output, such as the contents of some variables. But please note that a core file is only useful if paired with the exact same binary as generated the corefile. If you recompile Squid then any coredumps from previous versions will be useless unless you have saved the corresponding Squid binaries, and any attempts to analyze such coredumps will most certainly give misleading information about the cause to the crash.
If you CANNOT get Squid to leave a core file for you then one of the following approaches can be used
First alternative is to start Squid under the contol of GDB
% gdb /path/to/squid handle SIGPIPE pass nostop noprint run -DNYCd3 [wait for crash] backtrace quit
The drawback from the above is that it isn't really suitable to run on a production system as Squid then won't restart automatically if it crashes. The good news is that it is fully possible to automate the process above to automatically get the stack trace and then restart Squid. Here is a short automated script that should work:
#!/bin/sh
trap "rm -f $$.gdb" 0
cat <<EOF >$$.gdb
handle SIGPIPE pass nostop noprint
run -DNYCd3
backtrace
quit
EOF
while sleep 2; do
gdb -x $$.gdb /path/to/squid 2>&1 | tee -a squid.out
done
Other options if the above cannot be done is to:
a) Build Squid with the --enable-stacktraces option, if support exists for your OS (exists for Linux glibc on Intel, and Solaris with some extra libraries which seems rather impossible to find these days..)
b) Run Squid using the "catchsegv" tool. (Linux glibc Intel)
but these approaches does not by far provide as much details as using gdb.
If you believe you have found a non-fatal bug (such as incorrect HTTP processing) please send us a section of your cache.log with debugging to demonstrate the problem. The cache.log file can become very large, so alternatively, you may want to copy it to an FTP or HTTP server where we can download it.
It is very simple to enable full debugging on a running squid process. Simply use the -k debug command line option:
% ./squid -k debugThis causes every debug() statement in the source code to write a line in the cache.log file. You also use the same command to restore Squid to normal debugging level.
To enable selective debugging (e.g. for one source file only), you need to edit squid.conf and add to the debug_options line. Every Squid source file is assigned a different debugging section. The debugging section assignments can be found by looking at the top of individual source files, or by reading the file doc/debug-levels.txt (correctly renamed to debug-sections.txt for Squid-2). You also specify the debugging level to control the amount of debugging. Higher levels result in more debugging messages. For example, to enable full debugging of Access Control functions, you would use
debug_options ALL,1 28,9Then you have to restart or reconfigure Squid.
Once you have the debugging captured to cache.log, take a look at it yourself and see if you can make sense of the behaviour which you see. If not, please feel free to send your debugging output to the squid-users or squid-bugs lists.
Squid normally tests your system's DNS configuration before it starts server requests. Squid tries to resolve some common DNS names, as defined in the dns_testnames configuration directive. If Squid cannot resolve these names, it could mean:
To disable this feature, use the -D command line option.
Note, Squid does NOT use the dnsservers to test the DNS. The test is performed internally, before the dnsservers start.
Starting with version 1.1.15, we have required that you first run
squid -zto create the swap directories on your filesystem. If you have set the cache_effective_user option, then the Squid process takes on the given userid before making the directories. If the cache_dir directory (e.g. /var/spool/cache) does not exist, and the Squid userid does not have permission to create it, then you will get the ``permission denied'' error. This can be simply fixed by manually creating the cache directory.
# mkdir /var/spool/cache # chown <userid> <groupid> /var/spool/cache # squid -z
Alternatively, if the directory already exists, then your operating system may be returning ``Permission Denied'' instead of ``File Exists'' on the mkdir() system call. This patch by Miquel van Smoorenburg should fix it.
Either (1) the Squid userid does not have permission to bind to the port, or (2) some other process has bound itself to the port. Remember that root privileges are required to open port numbers less than 1024. If you see this message when using a high port number, or even when starting Squid as root, then the port has already been opened by another process. Maybe you are running in the HTTP Accelerator mode and there is already a HTTP server running on port 80? If you're really stuck, install the way cool lsof utility to show you which process has your port in use.
This is explained in the Redirector section.
See the next question.
Note: The information here applies to version 2.2 and earlier.
Squid keeps an in-memory bitmap of disk files that are available for use, or are being used. The size of this bitmap is determined at run name, based on two things: the size of your cache, and the average (mean) cache object size.
The size of your cache is specified in squid.conf, on the cache_dir lines. The mean object size can also be specified in squid.conf, with the 'store_avg_object_size' directive. By default, Squid uses 13 Kbytes as the average size.
When allocating the bitmaps, Squid allocates this many bits:
2 * cache_size / store_avg_object_size
So, if you exactly specify the correct average object size, Squid should have 50% filemap bits free when the cache is full. You can see how many filemap bits are being used by looking at the 'storedir' cache manager page. It looks like this:
Store Directory #0: /usr/local/squid/cache First level subdirectories: 4 Second level subdirectories: 4 Maximum Size: 1024000 KB Current Size: 924837 KB Percent Used: 90.32% Filemap bits in use: 77308 of 157538 (49%) Flags:
Now, if you see the ``You've run out of swap file numbers'' message, then it means one of two things:
To check the average file size of object currently in your cache, look at the cache manager 'info' page, and you will find a line like:
Mean Object Size: 11.96 KB
To make the warning message go away, set 'store_avg_object_size' to that value (or lower) and then restart Squid.
Note: The information here is current for version 2.3
Calm down, this is now normal. Squid now dynamically allocates filemap bits based on the number of objects in your cache. You won't run out of them, we promise.
In Unix, things like processes and files have an owner. For Squid, the process owner and file owner should be the same. If they are not the same, you may get messages like ``permission denied.''
To find out who owns a file, use the ls -l command:
% ls -l /usr/local/squid/logs/access.log
A process is normally owned by the user who starts it. However, Unix sometimes allows a process to change its owner. If you specified a value for the effective_user option in squid.conf, then that will be the process owner. The files must be owned by this same userid.
If all this is confusing, then you probably should not be running Squid until you learn some more about Unix. As a reference, I suggest Learning the UNIX Operating System, 4th Edition.
If I try by way of a test, to access
ftp://username:password@ftpserver/somewhere/foo.tar.gzI get
somewhere/foo.tar.gz: Not a directory.
Use this URL instead:
ftp://username:password@ftpserver/%2fsomewhere/foo.tar.gz
This means your pinger program does not have root priveleges. You should either do this:
% su # make install-pingeror
# chown root /usr/local/squid/bin/pinger # chmod 4755 /usr/local/squid/bin/pinger
A forwarding loop is when a request passes through one proxy more than once. You can get a forwarding loop if
Forwarding loops are detected by examining the Via request header. Each cache which "touches" a request must add its hostname to the Via header. If a cache notices its own hostname in this header for an incoming request, it knows there is a forwarding loop somewhere.
NOTE: Squid may report a forwarding loop if a request goes through two caches that have the same visible_hostname value. If you want to have multiple machines with the same visible_hostname then you must give each machine a different unique_hostname so that forwarding loops are correctly detected.
When Squid detects a forwarding loop, it is logged to the cache.log file with the recieved Via header. From this header you can determine which cache (the last in the list) forwarded the request to you.
One way to reduce forwarding loops is to change a parent relationship to a sibling relationship.
Another way is to use cache_peer_access rules. For example:
# Our parent caches cache_peer A.example.com parent 3128 3130 cache_peer B.example.com parent 3128 3130 cache_peer C.example.com parent 3128 3130 # An ACL list acl PEERS src A.example.com acl PEERS src B.example.com acl PEERS src C.example.com # Prevent forwarding loops cache_peer_access A.example.com allow !PEERS cache_peer_access B.example.com allow !PEERS cache_peer_access C.example.com allow !PEERSThe above configuration instructs squid to NOT forward a request to parents A, B, or C when a request is received from any one of those caches.
This error message is seen mostly on Solaris systems. Mark Kennedy gives a great explanation:
Error 71 [EPROTO] is an obscure way of reporting that clients made it onto your server's TCP incoming connection queue but the client tore down the connection before the server could accept it. I.e. your server ignored its clients for too long. We've seen this happen when we ran out of file descriptors. I guess it could also happen if something made squid block for a long time.
Got these messages in my cache log - I guess it means that the index contents do not match the contents on disk.
1998/09/23 09:31:30| storeSwapInFileOpened: /var/cache/00/00/00000015: Size mismatch: 776(fstat) != 3785(object) 1998/09/23 09:31:31| storeSwapInFileOpened: /var/cache/00/00/00000017: Size mismatch: 2571(fstat) != 4159(object)
What does Squid do in this case?
NOTE, these messages are specific to Squid-2. These happen when Squid reads an object from disk for a cache hit. After it opens the file, Squid checks to see if the size is what it expects it should be. If the size doesn't match, the error is printed. In this case, Squid does not send the wrong object to the client. It will re-fetch the object from the source.
These messages are caused by buggy clients, mostly Netscape Navigator. What happens is, Netscape sends an HTTPS/SSL request over a persistent HTTP connection. Normally, when Squid gets an SSL request, it looks like this:
CONNECT www.buy.com:443 HTTP/1.0Then Squid opens a TCP connection to the destination host and port, and the real request is sent encrypted over this connection. Thats the whole point of SSL, that all of the information must be sent encrypted.
With this client bug, however, Squid receives a request like this:
GET https://www.buy.com/corp/ordertracking.asp HTTP/1.0 Accept: */* User-agent: Netscape ... ...Now, all of the headers, and the message body have been sent, unencrypted to Squid. There is no way for Squid to somehow turn this into an SSL request. The only thing we can do is return the error message.
Note, this browser bug does represent a security risk because the browser is sending sensitive information unencrypted over the network.
by Dave J Woolley (DJW at bts dot co dot uk)
These are illegal URLs, generally only used by illegal sites; typically the web site that supports a spammer and is expected to survive a few hours longer than the spamming account.
Their intention is to:
Any browser or proxy that works with them should be considered a security risk.
RFC 1738 has this to say about the hostname part of a URL:
The fully qualified domain name of a network host, or its IP address as a set of four decimal digit groups separated by ".". Fully qualified domain names take the form as described in Section 3.5 of RFC 1034 [13] and Section 2.1 of RFC 1123 [5]: a sequence of domain labels separated by ".", each domain label starting and ending with an alphanumerical character and possibly also containing "-" characters. The rightmost domain label will never start with a digit, though, which syntactically distinguishes all domain names from the IP addresses.
Whitespace characters (space, tab, newline, carriage return) are not allowed in URI's and URL's. Unfortunately, a number of Web services generate URL's with whitespace. Of course your favorite browser silently accomodates these bad URL's. The servers (or people) that generate these URL's are in violation of Internet standards. The whitespace characters should be encoded.
If you want Squid to accept URL's with whitespace, you have to decide how to handle them. There are four choices that you can set with the uri_whitespace option:
This likely means that your system does not have a loopback network device, or that device is not properly configured. All Unix systems should have a network device named lo0, and it should be configured with the address 127.0.0.1. If not, you may get the above error message. To check your system, run:
% ifconfig lo0The result should look something like:
lo0: flags=8049<UP,LOOPBACK,RUNNING,MULTICAST> mtu 16384
inet 127.0.0.1 netmask 0xff000000
If you use FreeBSD, see this.
The format of the cache_dir option changed with version
2.3. It now takes a type argument. All you need to do
is insert ufs in the line, like this:
cache_dir ufs /var/squid/cache ...
As of Squid 2.3, the default is to use internal DNS lookup code. The cache_dns_program and dns_children options are not known squid.conf directives in this case. Simply comment out these two options.
If you want to use external DNS lookups, with the dnsserver program, then add this to your configure command:
--disable-internal-dns
Sort of. As of Squid 2.3, the default is to use internal DNS lookup code. The dns_defnames option is only used with the external dnsserver processes. If you relied on dns_defnames before, you have three choices:
search
and domain lines from /etc/resolv.conf.``Connection reset by peer'' is an error code that Unix operating systems sometimes return for read, write, connect, and other system calls.
Connection reset means that the other host, the peer, sent us a RESET packet on a TCP connection. A host sends a RESET when it receives an unexpected packet for a nonexistent connection. For example, if one side sends data at the same time that the other side closes a connection, when the other side receives the data it may send a reset back.
The fact that these messages appear in Squid's log might indicate a problem, such as a broken origin server or parent cache. On the other hand, they might be ``normal,'' especially since some applications are known to force connection resets rather than a proper close.
You probably don't need to worry about them, unless you receive a lot of user complaints relating to SSL sites.
Rick Jones notes that if the server is running a Microsoft TCP stack, clients receive RST segments whenever the listen queue overflows. In other words, if the server is really busy, new connections receive the reset message. This is contrary to rational behaviour, but is unlikely to change.
This is an error message, generated by your operating system, in response to a connect() system call. It happens when there is no server at the other end listening on the port number that we tried to connect to.
Its quite easy to generate this error on your own. Simply telnet to a random, high numbered port:
% telnet localhost 12345 Trying 127.0.0.1... telnet: Unable to connect to remote host: Connection refusedIt happens because there is no server listening for connections on port 12345.
When you see this in response to a URL request, it probably means the origin server web site is temporarily down. It may also mean that your parent cache is down, if you have one.
You may get this message when you run commands like squid -krotate.
This error message usually means that the squid.pid file is missing. Since the PID file is normally present when squid is running, the absence of the PID file usually means Squid is not running. If you accidentally delete the PID file, Squid will continue running, and you won't be able to send it any signals.
If you accidentally removed the PID file, there are two ways to get it back.
ps and find the Squid process id. You'll probably see
two processes, like this:
bender-wessels % ps ax | grep squid 83617 ?? Ss 0:00.00 squid -s 83619 ?? S 0:00.48 (squid) -s (squid)You want the second process id, 83619 in this case. Create the PID file and put the process id number there. For example:
echo 83619 > /usr/local/squid/logs/squid.pid
squid -kreconfigure:
kill -HUP 83619The reconfigure process creates a new PID file automatically.
You are probably starting Squid as root. Squid is trying to find a group-id that doesn't have any special priveleges that it will run as. The default is nogroup, but this may not be defined on your system. You need to edit squid.conf and set cache_effective_group to the name of an unpriveledged group from /etc/group. There is a good chance that nobody will work for you.
Note: The information here is current for version 2.3.
This is correct. Squid does not know what to do with an https URL. To handle such a URL, Squid would need to speak the SSL protocol. Unfortunately, it does not (yet).
Normally, when you type an https URL into your browser, one of two things happens.
The CONNECT method is a way to tunnel any kind of connection through an HTTP proxy. The proxy doesn't understand or interpret the contents. It just passes bytes back and forth between the client and server. For the gory details on tunnelling and the CONNECT method, please see RFC 2817 and Tunneling TCP based protocols through Web proxy servers (expired).
There may be many causes for this.
Andrew Doroshenko reports that removing /dev/null, or mounting a filesystem with the nodev option, can cause Squid to use 100% of CPU. His suggested solution is to ``touch /dev/null.''
Mikael Andersson reports that clicking on Webmin's cachemgr.cgi link creates numerous instances of cachemgr.cgi that quickly consume all available memory and brings the system to its knees.
Joe Cooper reports this to be caused by SSL problems in some browsers (mainly Netscape 6.x/Mozilla) if your Webmin is SSL enabled. Try with another browser such as Netscape 4.x or Microsoft IE, or disable SSL encryption in Webmin.
Some versions of GCC (notably 2.95.1 through 2.95.4 at least) have bugs with compiler optimization. These GCC bugs may cause NULL pointer accesses in Squid, resulting in a ``FATAL: Received Segment Violation...dying'' message and a core dump.
You can work around these GCC bugs by disabling compiler optimization. The best way to do that is start with a clean source tree and set the CC options specifically:
% cd squid-x.y % make distclean % setenv CFLAGS='-g -Wall' % ./configure ...
To check that you did it right, you can search for AC_CFLAGS in src/Makefile:
% grep AC_CFLAGS src/Makefile AC_CFLAGS = -g -WallNow when you recompile, GCC won't try to optimize anything:
% make Making all in lib... gcc -g -Wall -I../include -I../include -c rfc1123.c ...etc...
NOTE: some people worry that disabling compiler optimization will negatively impact Squid's performance. The impact should be negligible, unless your cache is really busy and already runs at a high CPU usage. For most people, the compiler optimization makes little or no difference at all.
By Yomler of fnac.net
A combination of a bad configuration of Internet Explorer and any application which use the cydoor DLLs will produce the entry in the log. See cydoor.com for a complete list.
The bad configuration of IE is the use of a active configuration script (proxy.pac) and an active or inactive, but filled proxy settings. IE will only use the proxy.pac. Cydoor aps will use both and will generate the errors.
Disabling the old proxy settings in IE is not enought, you should delete them completely and only use the proxy.pac for example.
By Henrik Nordström
Some people have asked why requests for domain names using national symbols as "supported" by the certain domain registrars does not work in Squid. This is because there as of yet is no standard on how to manage national characters in the current Internet protocols such as HTTP or DNS. The current Internet standards is very strict on what is an acceptable hostname and only accepts A-Z a-z 0-9 and - in Internet hostname labels. Anything outside this is outside the current Internet standards and will cause interoperability issues such as the problems seen with such names and Squid.
When there is a consensus in the DNS and HTTP standardization groups on how to handle international domain names Squid will be changed to support this if any changes to Squid will be required.
If you are interested in the progress of the standardization process for international domain names please see the IETF IDN working group's dedicated page.
This happens when Squid makes a TCP connection to an origin server, but for some reason, the connection is closed before Squid reads any data. Depending on various factors, Squid may be able to retry the request again. If you see the ``Zero Sized Reply'' error message, it means that Squid was unable to retry, or that all retry attempts also failed.
What causes a connection to close prematurely? It could be a number of things, including:
You may be able to use tcpdump to track down and observe the problem.
Some users believe the problem is caused by very large cookies. One user reports that his Zero Sized Reply problem went away when he told Internet Explorer to not accept third-party cookies.
Here are some things you can try to reduce the occurance of the Zero Sized Reply error:
echo 0 > /proc/sys/net/ipv4/tcp_ecn/.If this error causes serious problems for you and the above does not help, Squid developers would be happy to help you uncover the problem. However, we will require high-quality debugging information from you, such as tcpdump output, server IP addresses, operating system versions, and access.log entries with full HTTP headers.
If you want to make Squid give the Zero Sized error on demand, you can use the short C program below. Simply compile and start the program on a system that doesn't already have a server running on port 80. Then try to connect to this fake server through Squid:
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <assert.h>
int
main(int a, char **b)
{
struct sockaddr_in S;
int s,t,x;
s = socket(PF_INET, SOCK_STREAM, 0);
assert(s > 0);
memset(&S, '\0', sizeof(S));
S.sin_family = AF_INET;
S.sin_port = htons(80);
x = bind(s, (struct sockaddr *) &S, sizeof(S));
assert(x == 0);
x = listen(s, 10);
assert(x == 0);
while (1) {
struct sockaddr_in F;
int fl = sizeof(F);
t = accept(s, (struct sockaddr *) &F, &fl);
fprintf(stderr, "accpeted FD %d from %s:%d\n",
t, inet_ntoa(F.sin_addr), (int)ntohs(F.sin_port));
close(t);
fprintf(stderr, "closed FD %d\n", t);
}
return 0;
}
by Grzegorz Janoszka
This error message appears when you try downloading large file using GET or uploading it using POST/PUT. There are three parameters to look for: request_body_max_size, reply_body_max_size (these two are set to 0 by default now, which means no limits at all, earlier version of squid had e.g. 1MB in request) and request_header_max_size - it defaults to 10kB (now, earlier versions had here 4 or even 2 kB) - in some rather rare circumstances even 10kB is too low, so you can increase this value.
In some situations where swap.state has been corrupted Squid can be very confused about how much data it has in the cache. Such corruption may happen after a power failure or similar fatal event. To recover first stop Squid, then delete the swap.state files from each cache directory and then start Squid again. Squid will automatically rebuild the swap.state index from the cached files reasonably well.
If this does not work or causes too high load on your server due to the reindexing of the cache then delete the cache content as explained in I want to restart Squid with a clean cache.
By Janno de Wit
There seems to be some problems with Microsoft Windows to access the Windows Update website. This is especially a problem when you block all traffic by a firewall and force your users to go through the Squid Cache.
Symptom: WindowsUpdate gives error codes like 0x80072EFD and cannot update, automatic updates aren't working too.
Cause: In earlier Windows-versions WindowsUpdate takes the proxy-settings from Internet Explorer. Since XP SP2 this is not sure. At my machine I ran Windows XP SP1 without WindowsUpdate problems. When I upgraded to SP2 WindowsUpdate started to give errors when searching updates etc.
The problem was that WU did not go through the proxy and tries to establish direct HTTP connections to Update-servers. Even when I set the proxy in IE again, it didn't help . It isn't Squid's problem that Windows Update doesn't work, but it is in Windows itself. The solution is to use the 'proxycfg' tool shipped with Windows XP. With this tool you can set the proxy for WinHTTP.
Commands:
C:\> proxycfg # gives information about the current connection type. Note: 'Direct Connection' does not force WU to bypass proxy C:\> proxycfg -d # Set Direct Connection C:\> proxycfg -p wu-proxy.lan:8080 # Set Proxy to use with Windows Update to wu-proxy.lan, port 8080 c:\> proxycfg -u # Set proxy to Internet Explorer settings.
An Internet Object is a file, document or response to a query for an Internet service such as FTP, HTTP, or gopher. A client requests an Internet object from a caching proxy; if the object is not already cached, the proxy server fetches the object (either from the host specified in the URL or from a parent or sibling cache) and delivers it to the client.
ICP is a protocol used for communication among squid caches. The ICP protocol is defined in two Internet RFC's. RFC 2186 describes the protocol itself, while RFC 2187 describes the application of ICP to hierarchical Web caching.
ICP is primarily used within a cache hierarchy to locate specific objects in sibling caches. If a squid cache does not have a requested document, it sends an ICP query to its siblings, and the siblings respond with ICP replies indicating a ``HIT'' or a ``MISS.'' The cache then uses the replies to choose from which cache to resolve its own MISS.
ICP also supports multiplexed transmission of multiple object streams over a single TCP connection. ICP is currently implemented on top of UDP. Current versions of Squid also support ICP via multicast.
The dnsserver is a process forked by squid to
resolve IP addresses from domain names. This is necessary because
the gethostbyname(3) function blocks the calling process
until the DNS query is completed.
Squid must use non-blocking I/O at all times, so DNS lookups are implemented external to the main process. The dnsserver processes do not cache DNS lookups, that is implemented inside the squid process.
ftpget exists only in Squid 1.1 and Squid 1.0 versions.
The ftpget program is an FTP client used for retrieving files from FTP servers. Because the FTP protocol is complicated, it is easier to implement it separately from the main squid code.
FTP PUT should work with Squid-2.0 and later versions. If you are using Squid-1.1, then you need to upgrade before PUT will work.
A cache hierarchy is a collection of caching proxy servers organized in a logical parent/child and sibling arrangement so that caches closest to Internet gateways (closest to the backbone transit entry-points) act as parents to caches at locations farther from the backbone. The parent caches resolve ``misses'' for their children. In other words, when a cache requests an object from its parent, and the parent does not have the object in its cache, the parent fetches the object, caches it, and delivers it to the child. This ensures that the hierarchy achieves the maximum reduction in bandwidth utilization on the backbone transit links, helps reduce load on Internet information servers outside the network served by the hierarchy, and builds a rich cache on the parents so that the other child caches in the hierarchy will obtain better ``hit'' rates against their parents.
In addition to the parent-child relationships, squid supports the notion of siblings: caches at the same level in the hierarchy, provided to distribute cache server load. Each cache in the hierarchy independently decides whether to fetch the reference from the object's home site or from parent or sibling caches, using a a simple resolution protocol. Siblings will not fetch an object for another sibling to resolve a cache ``miss.''
The algorithm is somewhat more complicated when firewalls are involved.
The single_parent_bypass directive can be used to skip
the ICP queries if the only appropriate sibling is a parent cache
(i.e., if there's only one place you'd fetch the object from, why
bother querying?)
There are several open issues for the caching project namely more automatic load balancing and (both configured and dynamic) selection of parents, routing, multicast cache-to-cache communication, and better recognition of URLs that are not worth caching.
For our other to-do list items, please see our ``TODO'' file in the recent source distributions.
Prospective developers should review the resources available at the Squid developers corner
Workload can be characterized as the burden a client or group of clients imposes on a system. Understanding the nature of workloads is important to the managing system capacity.
If you are interested in Internet traffic workloads then NLANR's Network Analysis activities is a good place to start.
The NLANR root caches are at the NSF supercomputer centers (SCCs), which are interconnected via NSF's high speed backbone service (vBNS). So inter-cache communication between the NLANR root caches does not cross the Internet.
The benefits of hierarchical caching (namely, reduced network bandwidth consumption, reduced access latency, and improved resiliency) come at a price. Caches higher in the hierarchy must field the misses of their descendents. If the equilibrium hit rate of a leaf cache is 50%, half of all leaf references have to be resolved through a second level cache rather than directly from the object's source. If this second level cache has most of the documents, it is usually still a win, but if higher level caches often don't have the document, or become overloaded, then they could actually increase access latency, rather than reduce it.
Please see the Firewalls FAQ information site.
For example:
Storage LRU Expiration Age: 4.31 days
The LRU expiration age is a dynamically-calculated value. Any objects which have not been accessed for this amount of time will be removed from the cache to make room for new, incoming objects. Another way of looking at this is that it would take your cache approximately this many days to go from empty to full at your current traffic levels.
As your cache becomes more busy, the LRU age becomes lower so that more objects will be removed to make room for the new ones. Ideally, your cache will have an LRU age value in the range of at least 3 days. If the LRU age is lower than 3 days, then your cache is probably not big enough to handle the volume of requests it receives. By adding more disk space you could increase your cache hit ratio.
The configuration parameter reference_age places an upper limit on your cache's LRU expiration age.
Consider a pair of caches named A and B. It may be the case that A can reach B, and vice-versa, but B has poor reachability to the rest of the Internet. In this case, we would like B to recognize that it has poor reachability and somehow convey this fact to its neighbor caches.
Squid will track the ratio of failed-to-successful requests over short time periods. A failed request is one which is logged as ERR_DNS_FAIL, ERR_CONNECT_FAIL, or ERR_READ_ERROR. When the failed-to-successful ratio exceeds 1.0, then Squid will return ICP_MISS_NOFETCH instead of ICP_MISS to neighbors. Note, Squid will still return ICP_HIT for cache hits.
No, you must send a HUP signal to have Squid re-read its configuration file, including access control lists. An easy way to do this is with the -k command line option:
squid -k reconfigure
unlinkd is an external process used for unlinking unused cache files. Performing the unlink operation in an external process opens up some race-condition problems for Squid. If we are not careful, the following sequence of events could occur:
So, the problem is, how can we guarantee that unlinkd will not remove a cache file that Squid has recently allocated to a new object? The approach we have taken is to have Squid keep a stack of unused (but not deleted!) swap file numbers. The stack size is hard-coded at 128 entries. We only give unlink requests to unlinkd when the unused file number stack is full. Thus, if we ever have to start unlinking files, we have a pool of 128 file numbers to choose from which we know will not be removed by unlinkd.
In terms of implementation, the only way to send unlink requests to the unlinkd process is via the storePutUnusedFileno function.
Unfortunately there are times when Squid can not use the unlinkd process but must call unlink(2) directly. One of these times is when the cache swap size is over the high water mark. If we push the released file numbers onto the unused file number stack, and the stack is not full, then no files will be deleted, and the actual disk usage will remain unchanged. So, when we exceed the high water mark, we must call unlink(2) directly.
One of the most unpleasant things Squid must do is generate HTML pages of Gopher and FTP directory listings. For some strange reason, people like to have little icons next to each listing entry, denoting the type of object to which the link refers (image, text file, etc.).
In Squid 1.0 and 1.1, we used internal browser icons with names like gopher-internal-image. Unfortunately, these were not very portable. Not all browsers had internal icons, or even used the same names. Perhaps only Netscape and Mosaic used these names.
For Squid 2 we include a set of icons in the source distribution. These icon files are loaded by Squid as cached objects at runtime. Thus, every Squid cache now has its own icons to use in Gopher and FTP listings. Just like other objects available on the web, we refer to the icons with Uniform Resource Locators, or URLs.
Nope, its not possible. Squid only accepts HTTP requests. It speaks FTP on the server-side, but not on the client-side.
The very cool wget will download FTP URLs via Squid (and probably any other proxy cache).
Is there any way to speed up the time spent dealing with select? Cachemgr shows:
Select loop called: 885025 times, 714.176 ms avg
This number is NOT how much time it takes to handle filedescriptor I/O. We simply count the number of times select was called, and divide the total process running time by the number of select calls.
This means, on average it takes your cache .714 seconds to check all the open file descriptors once. But this also includes time select() spends in a wait state when there is no I/O on any file descriptors. My relatively idle workstation cache has similar numbers:
Select loop called: 336782 times, 715.938 ms avg
But my busy caches have much lower times:
Select loop called: 16940436 times, 10.427 ms avg
Select loop called: 80524058 times, 10.030 ms avg
Select loop called: 10590369 times, 8.675 ms avg
Select loop called: 84319441 times, 9.578 ms avg
The presence of Cookies headers in requests does not affect whether or not an HTTP reply can be cached. Similarly, the presense of Set-Cookie headers in replies does not affect whether the reply can be cached.
The proper way to deal with Set-Cookie reply headers, according to RFC 2109 is to cache the whole object, EXCEPT the Set-Cookie header lines.
With Squid-1.1, we can not filter out specific HTTP headers, so Squid-1.1 does not cache any response which contains a Set-Cookie header.
With Squid-2, however, we can filter out specific HTTP headers. But instead of filtering them on the receiving-side, we filter them on the sending-side. Thus, Squid-2 does cache replies with Set-Cookie headers, but it filters out the Set-Cookie header itself for cache hits.
When checking the object freshness, we calculate these values:
OBJ_AGE = NOW - OBJ_DATE
LM_AGE = OBJ_DATE - OBJ_LASTMOD
LM_FACTOR = OBJ_AGE / LM_AGE
These values are compared with the parameters of the refresh_pattern rules. The refresh parameters are:
The URL regular expressions are checked in the order listed until a match is found. Then the algorithms below are applied for determining if an object is fresh or stale.
if (CLIENT_MAX_AGE)
if (OBJ_AGE > CLIENT_MAX_AGE)
return STALE
if (OBJ_AGE <= CONF_MIN)
return FRESH
if (EXPIRES) {
if (EXPIRES <= NOW)
return STALE
else
return FRESH
}
if (OBJ_AGE > CONF_MAX)
return STALE
if (LM_FACTOR < CONF_PERCENT)
return FRESH
return STALE
Kolics Bertold has made an excellent flow chart diagram showing this process.
For Squid-2 the refresh algorithm has been slightly modified to give the EXPIRES value a higher precedence, and the CONF_MIN value lower precedence:
if (EXPIRES) {
if (EXPIRES <= NOW)
return STALE
else
return FRESH
}
if (CLIENT_MAX_AGE)
if (OBJ_AGE > CLIENT_MAX_AGE)
return STALE
if (OBJ_AGE > CONF_MAX)
return STALE
if (OBJ_DATE > OBJ_LASTMOD) {
if (LM_FACTOR < CONF_PERCENT)
return FRESH
else
return STALE
}
if (OBJ_AGE <= CONF_MIN)
return FRESH
return STALE
The cachemanager I/O page lists deferred reads for various server-side protocols.
Sometimes reading on the server-side gets ahead of writing to the client-side. Especially if your cache is on a fast network and your clients are connected at modem speeds. Squid-1.1 will read up to 256k (per request) ahead before it starts to defer the server-side reads.
I've been monitoring the traffic on my cache's ethernet adapter an found a behavior I can't explain: the inbound traffic is equal to the outbound traffic. The differences are negligible. The hit ratio reports 40%. Shouldn't the outbound be at least 40% greater than the inbound?
I can't account for the exact behavior you're seeing, but I can offer this advice; whenever you start measuring raw Ethernet or IP traffic on interfaces, you can forget about getting all the numbers to exactly match what Squid reports as the amount of traffic it has sent/received.
Why?
Squid is an application - it counts whatever data is sent to, or received from, the lower-level networking functions; at each successively lower layer, additional traffic is involved (such as header overhead, retransmits and fragmentation, unrelated broadcasts/traffic, etc.). The additional traffic is never seen by Squid and thus isn't counted - but if you run MRTG (or any SNMP/RMON measurement tool) against a specific interface, all this additional traffic will "magically appear".
Also remember that an interface has no concept of upper-layer networking (so an Ethernet interface doesn't distinguish between IP traffic that's entirely internal to your organization, and traffic that's to/from the Internet); this means that when you start measuring an interface, you have to be aware of *what* you are measuring before you can start comparing numbers elsewhere.
It is possible (though by no means guaranteed) that you are seeing roughly equivalent input/output because you're measuring an interface that both retrieves data from the outside world (Internet), *and* serves it to end users (internal clients). That wouldn't be the whole answer, but hopefully it gives you a few ideas to start applying to your own circumstance.
To interpret any statistic, you have to first know what you are measuring; for example, an interface counts inbound and outbound bytes - that's it. The interface doesn't distinguish between inbound bytes from external Internet sites or from internal (to the organization) clients (making requests). If you want that, try looking at RMON2.
Also, if you're talking about a 40% hit rate in terms of object requests/counts then there's absolutely no reason why you should expect a 40% reduction in traffic; after all, not every request/object is going to be the same size so you may be saving a lot in terms of requests but very little in terms of actual traffic.
To determine whether a given object may be cached, Squid takes many things into consideration. The current algorithm (for Squid-2) goes something like this:
The keep-alive ratio shows up in the server_list cache manager page for Squid 2.
This is a mechanism to try detecting neighbor caches which might not be able to deal with persistent connections. Every time we send a proxy-connection: keep-alive request header to a neighbor, we count how many times the neighbor sent us a proxy-connection: keep-alive reply header. Thus, the keep-alive ratio is the ratio of these two counters.
If the ratio stays above 0.5, then we continue to assume the neighbor properly implements persistent connections. Otherwise, we will stop sending the keep-alive request header to that neighbor.
Squid uses an LRU (least recently used) algorithm to replace old cache objects. This means objects which have not been accessed for the longest time are removed first. In the source code, the StoreEntry->lastref value is updated every time an object is accessed.
Objects are not necessarily removed ``on-demand.'' Instead, a regularly scheduled event runs to periodically remove objects. Normally this event runs every second.
Squid keeps the cache disk usage between the low and high water marks. By default the low mark is 90%, and the high mark is 95% of the total configured cache size. When the disk usage is close to the low mark, the replacement is less aggressive (fewer objects removed). When the usage is close to the high mark, the replacement is more aggressive (more objects removed).
When selecting objects for removal, Squid examines some number of objects and determines which can be removed and which cannot. A number of factors determine whether or not any given object can be removed. If the object is currently being requested, or retrieved from an upstream site, it will not be removed. If the object is ``negatively-cached'' it will be removed. If the object has a private cache key, it will be removed (there would be no reason to keep it -- because the key is private, it can never be ``found'' by subsequent requests). Finally, if the time since last access is greater than the LRU threshold, the object is removed.
The LRU threshold value is dynamically calculated based on the current cache size and the low and high marks. The LRU threshold scaled exponentially between the high and low water marks. When the store swap size is near the low water mark, the LRU threshold is large. When the store swap size is near the high water mark, the LRU threshold is small. The threshold automatically adjusts to the rate of incoming requests. In fact, when your cache size has stabilized, the LRU threshold represents how long it takes to fill (or fully replace) your cache at the current request rate. Typical values for the LRU threshold are 1 to 10 days.
Back to selecting objects for removal. Obviously it is not possible to check every object in the cache every time we need to remove some of them. We can only check a small subset each time. The way in which this is implemented is very different between Squid-1.1 and Squid-2.
The Squid cache storage is implemented as a hash table with some number of "hash buckets." Squid-1.1 scans one bucket at a time and sorts all the objects in the bucket by their LRU age. Objects with an LRU age over the threshold are removed. The scan rate is adjusted so that it takes approximately 24 hours to scan the entire cache. The store buckets are randomized so that we don't always scan the same buckets at the same time of the day.
This algorithm has some flaws. Because we only scan one bucket, there are going to be better candidates for removal in some of the other 16,000 or so buckets. Also, the qsort() function might take a non-trivial amount of CPU time, depending on how many entries are in each bucket.
For Squid-2 we eliminated the need to use qsort() by indexing cached objects into an automatically sorted linked list. Every time an object is accessed, it gets moved to the top of the list. Over time, the least used objects migrate to the bottom of the list. When looking for objects to remove, we only need to check the last 100 or so objects in the list. Unfortunately this approach increases our memory usage because of the need to store three additional pointers per cache object. But for Squid-2 we're still ahead of the game because we also replaced plain-text cache keys with MD5 hashes.
keys refers to the database keys which Squid uses to index cache objects. Every object in the cache--whether saved on disk or currently being downloaded--has a cache key. For Squid-1.0 and Squid-1.1 the cache key was basically the URL. Squid-2 uses MD5 checksums for cache keys.
The Squid cache uses the notions of private and public cache keys. An object can start out as being private, but may later be changed to public status. Private objects are associated with only a single client whereas a public object may be sent to multiple clients at the same time. In other words, public objects can be located by any cache client. Private keys can only be located by a single client--the one who requested it.
Objects are changed from private to public after all of the HTTP reply headers have been received and parsed. In some cases, the reply headers will indicate the object should not be made public. For example, if the no-cache Cache-Control directive is used.
We use it to collect data for Plankton.
It may. This is an old feature from the Harvest cache software. The cache would send ICP ``SECHO'' message to the echo ports of origin servers. If the SECHO message came back before any of the other ICP replies, then it meant the origin server was probably closer than any neighbor cache. In that case Harvest/Squid sent the request directly to the origin server.
With more attention focused on security, many administrators filter UDP packets to port 7. The Computer Emergency Response Team (CERT) once issued an advisory note ( CA-96.01: UDP Port Denial-of-Service Attack) that says UDP echo and chargen services can be used for a denial of service attack. This made admins extremely nervous about any packets hitting port 7 on their systems, and they made complaints.
The source_ping feature has been disabled in Squid-2. If you're seeing packets to port 7 that are coming from a Squid cache (remote port 3130), then its probably a very old version of Squid.
It means Squid sent a DNS query to one IP address, but the response came back from a different IP address. By default Squid checks that the addresses match. If not, Squid ignores the response.
There are a number of reasons why this would happen:
If you recognize the IP address in the warning as one of your name server hosts, then its probably numbers (1) or (2).
You can make these warnings stop, and allow responses from ``unknown'' name servers by setting this configuration option:
ignore_unknown_nameservers off
Note: The information here is current for version 2.2.
See storeDirMapAllocate() in the source code.
When Squid wants to create a new disk file for storing an object, it first selects which cache_dir the object will go into. This is done with the storeDirSelectSwapDir() function. If you have N cache directories, the function identifies the 3N/4 (75%) of them with the most available space. These directories are then used, in order of having the most available space. When Squid has stored one URL to each of the 3N/4 cache_dir's, the process repeats and storeDirSelectSwapDir() finds a new set of 3N/4 cache directories with the most available space.
Once the cache_dir has been selected, the next step is to find an available swap file number. This is accomplished by checking the file map, with the file_map_allocate() function. Essentially the swap file numbers are allocated sequentially. For example, if the last number allocated happens to be 1000, then the next one will be the first number after 1000 that is not already being used.
Byte hit ratio is calculated a bit differently than Request hit ratio. Squid counts the number of bytes read from the network on the server-side, and the number of bytes written to the client-side. The byte hit ratio is calculated as
(client_bytes - server_bytes) / client_bytesIf server_bytes is greater than client_bytes, you end up with a negative value.
The server_bytes may be greater than client_bytes for a number of reasons, including:
First you need to understand the difference between public and private keys.
When Squid sends ICP queries, it uses the ICP reqnum field to hold the private key data. In other words, when Squid gets an ICP reply, it uses the reqnum value to build the private cache key for the pending object.
Some ICP implementations always set the reqnum field to zero when they send a reply. Squid can not use private cache keys with such neighbor caches because Squid will not be able to locate cache keys for those ICP replies. Thus, if Squid detects a neighbor cache that sends zero reqnum's, it disables the use of private cache keys.
Not having private cache keys has some important privacy implications. Two users could receive one response that was meant for only one of the users. This response could contain personal, confidential information. You will need to disable the ``zero reqnum'' neighbor if you want Squid to use private cache keys.
TCP allows connections to be in a ``half-closed'' state. This is accomplished with the shutdown(2) system call. In Squid, this means that a client has closed its side of the connection for writing, but leaves it open for reading. Half-closed connections are tricky because Squid can't tell the difference between a half-closed connection, and a fully closed one.
If Squid tries to read a connection, and read() returns 0, and Squid knows that the client doesn't have the whole response yet, Squid puts marks the filedescriptor as half-closed. Most likely the client has aborted the request and the connection is really closed. However, there is a slight chance that the client is using the shutdown() call, and that it can still read the response.
To disable half-closed connections, simply put this in squid.conf:
half_closed_clients offThen, Squid will always close its side of the connection instead of marking it as half-closed.
Squid has traditionally used an LRU replacement algorithm. As of version 2.3, you can use some other replacement algorithms by using the --enable-heap-replacement configure option. Currently, the heap replacement code supports two additional algorithms: LFUDA, and GDS.
With Squid version 2.4 and later you should use this configure option:
./configure --enable-removal-policies=heap
Then, in squid.conf, you can select different policies with the cache_replacement_policy option. See the squid.conf comments for details.
The LFUDA and GDS replacement code was contributed by John Dilley and others from Hewlett-Packard. Their work is described in these papers:
If you compare df output and cachemgr storedir output, you will notice that actual disk usage is greater than what Squid reports. This may be due to a number of reasons:
positive_dns_ttl is how long Squid caches a successful DNS lookup. Similarly, negative_dns_ttl is how long Squid caches a failed DNS lookup.
positive_dns_ttl is not always used. It is NOT used in the following cases:
Let's say you have the following settings:
positive_dns_ttl 1 hours negative_dns_ttl 1 minutesWhen Squid looks up a name like www.squid-cache.org, it gets back an IP address like 204.144.128.89. The address is cached for the next hour. That means, when Squid needs to know the address for www.squid-cache.org again, it uses the cached answer for the next hour. After one hour, the cached information expires, and Squid makes a new query for the address of www.squid-cache.org.
If you have the DNS TTL patch, or are using internal lookups, then each hostname has its own TTL value, which was set by the domain name administrator. You can see these values in the 'ipcache' cache manager page. For example:
Hostname Flags lstref TTL N www.squid-cache.org C 73043 12784 1( 0) 204.144.128.89-OK www.ircache.net C 73812 10891 1( 0) 192.52.106.12-OK polygraph.ircache.net C 241768 -181261 1( 0) 192.52.106.12-OKThe TTL field shows how how many seconds until the entry expires. Negative values mean the entry is already expired, and will be refreshed upon next use.
The negative_dns_ttl specifies how long to cache failed DNS lookups. When Squid fails to resolve a hostname, you can be pretty sure that it is a real failure, and you are not likely to get a successful answer within a short time period. Squid retries its lookups many times before declaring a lookup has failed. If you like, you can set negative_dns_ttl to zero.
It means that Squid opened up a disk file to serve a cache hit, but it found that the stored object doesn't match what the user's request. Squid stores the MD5 digest of the URL at the start of each disk file. When the file is opened, Squid checks that the disk file MD5 matches the MD5 of the URL requested by the user. If they don't match, the warning is printed and Squid forwards the request to the origin server.
You do not need to worry about this warning. It means that Squid is recovering from a corrupted cache directory.
Each of Squid's disk cache files has a metadata section at the beginning. This header is used to store the URL MD5, some StoreEntry data, and more. When Squid opens a disk file for reading, it looks for the meta data header and unpacks it.
This warning means that Squid couln't unpack the meta data. This is non-fatal bug, from which Squid can recover. Perhaps the meta data was just missing, or perhaps the file got corrupted.
You do not need to worry about this warning. It means that Squid is double-checking that the disk file matches what Squid thinks should be there, and the check failed. Squid recorvers and generates a cache miss in this case.
Its a side-effect of the way interception proxying works.
When Squid is configured for interception proxying, the operating system pretends that it is the origin server. That means that the "local" socket address for intercepted TCP connections is really the origin server's IP address. If you run netstat -n on your interception proxy, you'll see a lot of foreign IP addresses in the Local Address column.
When Squid wants to make an ident query, it creates a new TCP socket and binds the local endpoint to the same IP address as the local end of the client's TCP connection. Since the local address isn't really local (its some far away origin server's IP address), the bind() system call fails. Squid handles this as a failed ident lookup.
So why bind in that way? If you know you are interception proxying, then why not bind the local endpoint to the host's (intranet) IP address? Why make the masses suffer needlessly?
Because thats just how ident works. Please read RFC 931, in particular the RESTRICTIONS section.
This means that you are using external dnsserver processes for lookups, and all processes are busy, and Squid's pending queue is full. Each dnsserver program can only handle one request at a time. When all dnsserver processes are busy, Squid queues up requests, but only to a certain point.
To alleviate this condition, you need to either (1) increase the number of dnsserver processes by changing the value for dns_children in your config file, or (2) switch to using Squid's internal DNS client code.
Note that in some versions, Squid limits dns_children to 32. To increase it beyond that value, you would have to edit the source code.
by Colin Campbell
Ftp uses two data streams, one for passing commands around, the other for moving data. The command channel is handled by the ftpd listening on port 21.
The data channel varies depending on whether you ask for passive ftp or not. When you request data in a non-passive environment, you client tells the server ``I am listening on <ip-address> <port>.'' The server then connects FROM port 20 to the ip address and port specified by your client. This requires your "security device" to permit any host outside from port 20 to any host inside on any port > 1023. Somewhat of a hole.
In passive mode, when you request a data transfer, the server tells the client ``I am listening on <ip address> <port>.'' Your client then connects to the server on that IP and port and data flows.
Multicast is essentially the ability to send one IP packet to multiple receivers. Multicast is often used for audio and video conferencing systems.
One way is to ask someone who manages your network. If your network manager doesn't know, or looks at you funny, then you probably don't have it.
Another way is to use the mtrace program, which can be found on the Xerox PARC FTP site. Mtrace is similar to traceroute. It will tell you about the multicast path between your site and another. For example:
> mtrace mbone.ucar.edu
mtrace: WARNING: no multicast group specified, so no statistics printed
Mtrace from 128.117.64.29 to 192.172.226.25 via group 224.2.0.1
Querying full reverse path... * switching to hop-by-hop:
0 oceana-ether.nlanr.net (192.172.226.25)
-1 avidya-ether.nlanr.net (192.172.226.57) DVMRP thresh^ 1
-2 mbone.sdsc.edu (198.17.46.39) DVMRP thresh^ 1
-3 * nccosc-mbone.dren.net (138.18.5.224) DVMRP thresh^ 48
-4 * * FIXW-MBONE.NSN.NASA.GOV (192.203.230.243) PIM/Special thresh^ 64
-5 dec3800-2-fddi-0.SanFrancisco.mci.net (204.70.158.61) DVMRP thresh^ 64
-6 dec3800-2-fddi-0.Denver.mci.net (204.70.152.61) DVMRP thresh^ 1
-7 mbone.ucar.edu (192.52.106.7) DVMRP thresh^ 64
-8 mbone.ucar.edu (128.117.64.29)
Round trip time 196 ms; total ttl of 68 required.
Short answer: No, probably not.
Reasons why you SHOULD use Multicast:
Reasons why you SHOULD NOT use Multicast:
We only recommend people to use Multicast ICP over network infrastructure which they have close control over. In other words, only use Multicast over your local area network, or maybe your wide area network if you are an ISP. We think it is probably a bad idea to use Multicast ICP over congested links or commodity backbones.
To configure Squid to send ICP queries to a Multicast address, you need to create another neighbour cache entry specified as multicast. For example:
cache_peer 224.9.9.9 multicast 3128 3130 ttl=64224.9.9.9 is a sample multicast group address. multicast indicates that this is a special type of neighbour. The HTTP-port argument (3128) is ignored for multicast peers, but the ICP-port (3130) is very important. The final argument, ttl=64 specifies the multicast TTL value for queries sent to this address. It is probably a good idea to increment the minimum TTL by a few to provide a margin for error and changing conditions.
You must also specify which of your neighbours will respond to your multicast queries, since it would be a bad idea to implicitly trust any ICP reply from an unknown address. Note that ICP replies are sent back to unicast addresses; they are NOT multicast, so Squid has no indication whether a reply is from a regular query or a multicast query. To configure your multicast group neighbours, use the cache_peer directive and the multicast-responder option:
cache_peer cache1 sibling 3128 3130 multicast-responder cache_peer cache2 sibling 3128 3130 multicast-responderHere all fields are relevant. The ICP port number (3130) must be the same as in the cache_peer line defining the multicast peer above. The third field must either be parent or sibling to indicate how Squid should treat replies. With the multicast-responder flag set for a peer, Squid will NOT send ICP queries to it directly (i.e. unicast).
The Multicast TTL (which is specified on the cache_peer line of your multicast group) determines how ``far'' your ICP queries will go. In the Mbone, there is a certain TTL threshold defined for each network interface or tunnel. A multicast packet's TTL must be larger than the defined TTL for that packet to be forwarded across that link. For example, the mrouted manual page recommends:
32 for links that separate sites within an organization.
64 for links that separate communities or organizations, and are
attached to the Internet MBONE.
128 for links that separate continents on the MBONE.
A good way to determine the TTL you need is to run mtrace as shown above and look at the last line. It will show you the minimum TTL required to reach the other host.
If you set you TTL too high, then your ICP messages may travel ``too far'' and will be subject to eavesdropping by others. If you're only using multicast on your LAN, as we suggest, then your TTL will be quite small, for example ttl=4.
You must tell Squid to join a multicast group address with the mcast_groups directive. For example:
mcast_groups 224.9.9.9Of course, all members of your Multicast ICP group will need to use the exact same multicast group address.
NOTE: Choose a multicast group address with care! If two organizations happen to choose the same multicast address, then they may find that their groups ``overlap'' at some point. This will be especially true if one of the querying caches uses a large TTL value. There are two ways to reduce the risk of group overlap:
Using a unique address is a good idea, but not without some potential problems. If you choose an address randomly, how do you know that someone else will not also randomly choose the same address? NLANR has been assigned a block of multicast addresses by the IANA for use in situations such as this. If you would like to be assigned one of these addresses, please write to us. However, note that NLANR or IANA have no authority to prevent anyone from using an address assigned to you.
Limiting the scope of your multicast messages is probably a better solution. They can be limited with the TTL value discussed above, or with some newer techniques known as administratively scoped addresses. Here you can configure well-defined boundaries for the traffic to a specific address. The Administratively Scoped IP Multicast RFC describes this.
J.D. Bronson (jb at ktxg dot com) reported that his Solaris box could not talk to certain origin servers, such as moneycentral.msn.com and www.mbnanetaccess.com. J.D. fixed his problem by setting:
tcp_xmit_hiwat 49152 tcp_xmit_lowat 4096 tcp_recv_hiwat 49152
select(3c) won't handle more than 1024 file descriptors. The configure script should enable poll() by default for Solaris. poll() allows you to use many more filedescriptors, probably 8192 or more.
For older Squid versions you can enable poll() manually by changing HAVE_POLL in include/autoconf.h, or by adding -DUSE_POLL=1 to the DEFINES in src/Makefile.
libmalloc.a is leaky. Squid's configure does not use -lmalloc on Solaris.
by David J N Begley.
DNS lookups can be slow because of some mysterious thing called ncsd. You should edit /etc/nscd.conf and make it say:
enable-cache hosts no
Apparently nscd serializes DNS queries thus slowing everything down when an application (such as Squid) hits the resolver hard. You may notice something similar if you run a log processor executing many DNS resolver queries - the resolver starts to slow.. right.. down.. . . .
According to Andres Kroonmaa, users of Solaris starting from version 2.6 and up should NOT completely disable nscd daemon. nscd should be running and caching passwd and group files, although it is suggested to disable hosts caching as it may interfere with DNS lookups.
Several library calls rely on available free FILE descriptors FD < 256. Systems running without nscd may fail on such calls if first 256 files are all in use.
Since solaris 2.6 Sun has changed the way some system calls work and is using nscd daemon as a implementor of them. To communicate to nscd Solaris is using undocumented door calls. Basically nscd is used to reduce memory usage of user-space system libraries that use passwd and group files. Before 2.6 Solaris cached full passwd file in library memory on the first use but as this was considered to use up too much ram on large multiuser systems Sun has decided to move implementation of these calls out of libraries and to a single dedicated daemon.
by Jason Armistead.
The /etc/nsswitch.conf file determines the order of searches for lookups (amongst other things). You might only have it set up to allow NIS and HOSTS files to work. You definitely want the "hosts:" line to include the word dns, e.g.:
hosts: nis dns [NOTFOUND=return] files
by Chris Tilbury.
Our site cache is running on a Solaris 2.6 machine. We use NIS to distribute authentication and local hosts information around and in common with our multiuser systems, we run a slave NIS server on it to help the response of NIS queries.
We were seeing very high name-ip lookup times (avg ~2sec) and ip->name lookup times (avg ~8 sec), although there didn't seem to be that much of a problem with response times for valid sites until the cache was being placed under high load. Then, performance went down the toilet.
After some time, and a bit of detective work, we found the problem. On Solaris 2.6, if you have a local NIS server running (ypserv) and you have NIS in your /etc/nsswitch.conf hosts entry, then check the flags it is being started with. The 2.6 ypstart script checks to see if there is a resolv.conf file present when it starts ypserv. If there is, then it starts it with the -d option.
This has the same effect as putting the YP_INTERDOMAIN key in the hosts table -- namely, that failed NIS host lookups are tried against the DNS by the NIS server.
This is a bad thing(tm)! If NIS itself tries to resolve names using the DNS, then the requests are serialised through the NIS server, creating a bottleneck (This is the same basic problem that is seen with nscd). Thus, one failing or slow lookup can, if you have NIS before DNS in the service switch file (which is the most common setup), hold up every other lookup taking place.
If you're running in this kind of setup, then you will want to make sure that
We changed these here, and saw our average lookup times drop by up to an order of magnitude (~150msec for name-ip queries and ~1.5sec for ip-name queries, the latter still so high, I suspect, because more of these fail and timeout since they are not made so often and the entries are frequently non-existent anyway).
Solaris 2.x - tuning your TCP/IP stack and more by Jens-S. Vckler
You might get this error even if your disk is not full, and is not out of inodes. Check your syslog logs (/var/adm/messages, normally) for messages like either of these:
NOTICE: realloccg /proxy/cache: file system full NOTICE: alloc: /proxy/cache: file system full
In a nutshell, the UFS filesystem used by Solaris can't cope with the workload squid presents to it very well. The filesystem will end up becoming highly fragmented, until it reaches a point where there are insufficient free blocks left to create files with, and only fragments available. At this point, you'll get this error and squid will revise its idea of how much space is actually available to it. You can do a "fsck -n raw_device" (no need to unmount, this checks in read only mode) to look at the fragmentation level of the filesystem. It will probably be quite high (>15%).
Sun suggest two solutions to this problem. One costs money, the other is free but may result in a loss of performance (although Sun do claim it shouldn't, given the already highly random nature of squid disk access).
The first is to buy a copy of VxFS, the Veritas Filesystem. This is an extent-based filesystem and it's capable of having online defragmentation performed on mounted filesystems. This costs money, however (VxFS is not very cheap!)
The second is to change certain parameters of the UFS filesystem. Unmount your cache filesystems and use tunefs to change optimization to "space" and to reduce the "minfree" value to 3-5% (under Solaris 2.6 and higher, very large filesystems will almost certainly have a minfree of 2% already and you shouldn't increase this). You should be able to get fragmentation down to around 3% by doing this, with an accompanied increase in the amount of space available.
Thanks to Chris Tilbury.
by Jeff Madison
Important update regarding Squid running on Solaris x86. I have been working for several months to resolve what appeared to be a memory leak in squid when running on Solaris x86 regardless of the malloc that was used. I have made 2 discoveries that anyone running Squid on this platform may be interested in.
Number 1: There is not a memory leak in Squid even though after the system runs for some amount of time, this varies depending on the load the system is under, Top reports that there is very little memory free. True to the claims of the Sun engineer I spoke to this statistic from Top is incorrect. The odd thing is that you do begin to see performance suffer substantially as time goes on and the only way to correct the situation is to reboot the system. This leads me to discovery number 2.
Number 2: There is some type of resource problem, memory or other, with IPFilter on Solaris x86. I have not taken the time to investigate what the problem is because we no longer are using IPFilter. We have switched to a Alteon ACE 180 Gigabit switch which will do the trans-proxy for you. After moving the trans-proxy, redirection process out to the Alteon switch Squid has run for 3 days strait under a huge load with no problem what so ever. We currently have 2 boxes with 40 GB of cached objects on each box. This 40 GB was accumulated in the 3 days, from this you can see what type of load these boxes are under. Prior to this change we were never able to operate for more than 4 hours.
Because the problem appears to be with IPFilter I would guess that you would only run into this issue if you are trying to run Squid as a interception proxy using IPFilter. That makes sense. If there is anyone with information that would indicate my finding are incorrect I am willing to investigate further.
On Solaris, the kernel variable for the directory name lookup cache size is ncsize. In /etc/system, you might want to try
set ncsize = 8192or even higher. The kernel variable ufs_inode - which is the size of the inode cache itself - scales with ncsize in Solaris 2.5.1 and later. Previous versions of Solaris required both to be adjusted independently, but now, it is not recommended to adjust ufs_inode directly on 2.5.1 and later.
You can set ncsize quite high, but at some point - dependent on the application - a too-large ncsize will increase the latency of lookups.
Defaults are:
Solaris 2.5.1 : (max_nprocs + 16 + maxusers) + 64 Solaris 2.6/Solaris 7 : 4 * (max_nprocs + maxusers) + 320
Another new tuneable (actually a toggle) in Solaris 2.5.1, 2.6 or Solaris 7 is the priority_paging algorithm. This is actually a complete rewrite of the virtual memory system on Solaris. It will page out application data last, and filesystem pages first, if you turn it on (set priority_paging = 1 in /etc/system). As you may know, the Solaris buffer cache grows to fill available pages, and under the old VM system, applications could get paged out to make way for the buffer cache, which can lead to swap thrashing and degraded application performance. The new priority_paging helps keep application and shared library pages in memory, preventing the buffer cache from paging them out, until memory gets REALLY short. Solaris 2.5.1 requires patch 103640-25 or higher and Solaris 2.6 requires 105181-10 or higher to get priority_paging. Solaris 7 needs no patch, but all versions have it turned off by default.
by Marc
This crash happen on Solaris, when you don't have the "math.h" file at the compile time. I guess it can happen on every system without the correct include, but I have not verified.
The configure script just report: "math.h: no" and continue. The math functions are bad declared, and this cause this crash.
For 32bit Solaris, "math.h" is found in the SUNWlibm package.
We have found that with FreeBSD-2.2.2-RELEASE, there some bugs with T/TCP. FreeBSD will try to use T/TCP if you've enabled the ``TCP Extensions.'' To disable T/TCP, use sysinstall to disable TCP Extensions, or edit /etc/rc.conf and set
tcp_extensions="NO" # Allow RFC1323 & RFC1544 extensions (or NO).or add this to your /etc/rc files:
sysctl -w net.inet.tcp.rfc1644=0
We noticed an odd thing with some of Squid's interprocess communication. Often, output from the dnsserver processes would NOT be read in one chunk. With full debugging, it looks like this:
1998/04/02 15:18:48| comm_select: FD 46 ready for reading 1998/04/02 15:18:48| ipcache_dnsHandleRead: Result from DNS ID 2 (100 bytes) 1998/04/02 15:18:48| ipcache_dnsHandleRead: Incomplete reply ....other processing occurs... 1998/04/02 15:18:48| comm_select: FD 46 ready for reading 1998/04/02 15:18:48| ipcache_dnsHandleRead: Result from DNS ID 2 (9 bytes) 1998/04/02 15:18:48| ipcache_parsebuffer: parsing: $name www.karup.com $h_name www.karup.inter.net $h_len 4 $ipcount 2 38.15.68.128 38.15.67.128 $ttl 2348 $end
Interestingly, it is very common to get only 100 bytes on the first read. When two read() calls are required, this adds additional latency to the overall request. On our caches running Digital Unix, the median dnsserver response time was measured at 0.01 seconds. On our FreeBSD cache, however, the median latency was 0.10 seconds.
Here is a simple patch to fix the bug:
===================================================================
RCS file: /home/ncvs/src/sys/kern/uipc_socket.c,v
retrieving revision 1.40
retrieving revision 1.41
diff -p -u -r1.40 -r1.41
--- src/sys/kern/uipc_socket.c 1998/05/15 20:11:30 1.40
+++ /home/ncvs/src/sys/kern/uipc_socket.c 1998/07/06 19:27:14 1.41
@@ -31,7 +31,7 @@
* SUCH DAMAGE.
*
* @(#)uipc_socket.c 8.3 (Berkeley) 4/15/94
- * $Id: FAQ.sgml,v 1.250 2005/04/22 19:29:50 hno Exp $
+ * $Id: FAQ.sgml,v 1.250 2005/04/22 19:29:50 hno Exp $
*/
#include <sys/param.h>
@@ -491,6 +491,7 @@ restart:
mlen = MCLBYTES;
len = min(min(mlen, resid), space);
} else {
+ atomic = 1;
nopages:
len = min(min(mlen, resid), space);
/*
Another technique which may help, but does not fix the bug, is to increase the kernel's mbuf size. The default is 128 bytes. The MSIZE symbol is defined in /usr/include/machine/param.h. However, to change it we added this line to our kernel configuration file:
options MSIZE="256"
/var/yp/Makefile has the following section:
# The following line encodes the YP_INTERDOMAIN key into the hosts.byname # and hosts.byaddr maps so that ypserv(8) will do DNS lookups to resolve # hosts not in the current domain. Commenting this line out will disable # the DNS lookups. B=-bYou will want to comment out the B=-b line so that ypserv does not do DNS lookups.
Squid requires a the loopback interface to be up and configured. If it is not, you will get errors such as commBind.
From FreeBSD 3.3 Errata Notes:
Fix: Assuming that you experience this problem at all, edit /etc/rc.conf and search for where the network_interfaces variable is set. In its value, change the word auto to lo0 since the auto keyword doesn't bring the loop-back device up properly, for reasons yet to be adequately determined. Since your other interface(s) will already be set in the network_interfaces variable after initial installation, it's reasonable to simply s/auto/lo0/ in rc.conf and move on.
Thanks to Robert Lister.
FreeBSD 3.x and newer support Softupdates. This is a mechanism to speed up disk writes as it is possible by mounting ufs volumes async. However, Softupdates does this in a way that a performance similar or better than async is achieved but without loosing security in a case of a system crash. For more detailed information and the copyright terms see /sys/contrib/softupdates/README and /sys/ufs/ffs/README.softupdate.
To build a system supporting softupdates, you have to build
a kernel with options SOFTUPDATES set (see LINT for a commented
out example). After rebooting with the new kernel, you can enable
softupdates on a per filesystem base with the command:
$ tunefs -n /mountpointThe filesystem in question MUST NOT be mounted at this time. After that, softupdates are permanently enabled and the filesystem can be mounted normally. To verify that the softupdates code is running, simply issue a mount command and an output similar to the following will appear:
$ mount /dev/da2a on /usr/local/squid/cache (ufs, local, noatime, soft-updates, writes: sync 70 async 225)
Some users report problems with running Squid in the jail environment. Specifically, Squid logs messages like:
2001/10/12 02:08:49| comm_udp_sendto: FD 4, 192.168.1.3, port 53: (22) Invalid argument 2001/10/12 02:08:49| idnsSendQuery: FD 4: sendto: (22) Invalid argument
You can eliminate the problem by putting the jail's network interface address in the 'udp_outgoing_addr' configuration option in squid.conf.
On FreeBSD, make sure that TCP blackholing is not active. You can verify the current setting with:
# /sbin/sysctl net.inet.tcp.blackholeIt should return the following output:
net.inet.tcp.blackhole: 0If it is set to a positive value (usually, 2), disable it by setting it back to zero with<
# /sbin/sysctl net.inet.tcp.blackhole=0To make sure the setting survives across reboots, add the following line to the file /etc/sysctl.conf:
net.inet.tcp.blackhole=0
If you compile both libgnumalloc.a and Squid with cc, the mstats() function returns bogus values. However, if you compile libgnumalloc.a with gcc, and Squid with cc, the values are correct.
Some people report difficulties compiling squid on BSD/OS.
I've noticed that my Squid process seems to stick at a nice value of four, and clicks back to that even after I renice it to a higher priority. However, looking through the Squid source, I can't find any instance of a setpriority() call, or anything else that would seem to indicate Squid's adjusting its own priority.
by Bill Bogstad
BSD Unices traditionally have auto-niced non-root processes to 4 after they used alot (4 minutes???) of CPU time. My guess is that it's the BSD/OS not Squid that is doing this. I don't know offhand if there is a way to disable this on BSD/OS.
by Arjan de Vet
You can get around this by starting Squid with nice-level -4 (or another negative value).
The autonice behavior is a leftover from the history of BSD as a university OS. It penalises CPU bound jobs by nicing them after using 600 CPU seconds. Adding
sysctl -w kern.autonicetime=0to /etc/rc.local will disable the behavior systemwide.
Try a different version of Linux. We have received many reports of this ``bug'' from people running Linux 2.0.30. The bind(2) system call should NEVER give this error when binding to port 0.
Some users have reported that setting cache_effective_user
to nobody under Linux does not work.
However, it appears that using any cache_effective_user other
than nobody will succeed. One solution is to create a
user account for Squid and set cache_effective_user to that.
Alternately you can change the UID for the nobody account
from 65535 to 65534.
Another problem is that RedHat 5.0 Linux seems to have a broken setresuid() function. There are two ways to fix this. Before running configure:
% setenv ac_cv_func_setresuid no % ./configure ... % make clean % make installOr after running configure, manually edit include/autoconf.h and change the HAVE_SETRESUID line to:
#define HAVE_SETRESUID 0
Also, some users report this error is due to a NIS configuration problem. By adding compat to the passwd and group lines of /etc/nsswitch.conf, the problem goes away. ( Ambrose Li).
Russ Mellon notes that these problems with cache_effective_user are fixed in version 2.2.x of the Linux kernel.
The regular expression library which comes with Linux is known to be very slow. Some people report it entirely fails to work after long periods of time.
To fix, use the GNUregex library included with the Squid source code. With Squid-2, use the --enable-gnuregex configure option.
by Radu Greab
The gethostbyname() function leaks memory in RedHat 6.0 with glibc 2.1.1. The quick fix is to delete nisplus service from hosts entry in /etc/nsswitch.conf. In my tests dnsserver memory use remained stable after I made the above change.
See RedHat bug id 3919.
Some early versions of Linux have a kernel bug that causes this. All that is needed is a recent kernel that doesn't have the mentioned bug.
This is a bug with some versions of glibc. The glibc headers incorrectly depended on the contents of some kernel headers. Everything broke down when the kernel folks rearranged a bit in the kernel-specific header files.
We think this glibc bug is present in versions 2.1.1 (or 2.1.0) and earlier. There are two solutions:
When using Squid, some sites may give erorrs such as ``(111) Connection refused'' or ``(110) Connection timed out'' although these sites work fine without going through Squid.
Some versions of linux implement Explicit Congestion Notification (ECN) and this can cause some TCP connections to fail when contacting some sites with broken firewalls or broken TCP/IP implementations. A list of sites to be broken can be found at ECN Hall of Shame.
To work around such broken sites you can disable ECN with the following command:
echo 0 > /proc/sys/net/ipv4/tcp_ecn
Found this on the FreeBSD mailing list:
From: Robert Watson
As Bill Fumerola has indicated, and I thought I'd follow up in with a bit more detail, the behavior you're seeing is the result of a bug in the FreeBSD IPFW code. FreeBSD did a direct comparison of the TCP header flag field with an internal field in the IPFW rule description structure. Unfortunately, at some point, someone decided to overload the IPFW rule description structure field to add a flag representing "ESTABLISHED". They used a flag value that was previously unused by the TCP protocol (which doesn't make it safer, just less noticeable). Later, when that flag was allocated for ECN (Endpoint Congestion Notification) in TCP, and Linux began using ECN by default, the packets began to match ESTABLISHED rules regardless of the other TCP header flags. This bug was corrected on the RELENG_4 branch, and security advisory for the bug was released. This was, needless to say, a pretty serious bug, and good example of why you should be very careful to compare only the bits you really mean to, and should seperate packet state from protocol state in management structures, as well as make use of extensive testing to make sure rules actually have the effect you describe.
See also the thread on the NANOG mailing list, RFC3168 "The Addition of Explicit Congestion Notification (ECN) to IP, PROPOSED STANDARD" , Sally Floyd's page on ECN and problems related to it or ECN Hall of Shame for more information.
This was a very mysterious and unexplainable bug with GCC on HP-UX. Certain functions, when specified as static, would cause math bugs. The compiler also failed to handle implied int-double conversions properly. These bugs should all be handled correctly in Squid version 2.2.
There is a problem with GCC (2.8.1 at least) on Irix 6 which causes it to always return the string 255.255.255.255 for _ANY_ address when calling inet_ntoa(). If this happens to you, compile Squid with the native C compiler instead of GCC.
by F.J. Bosscha
To make squid run comfortable on SCO-unix you need to do the following:
Increase the NOFILES paramater and the NUMSP parameter and compile squid with I had, although squid told in the cache.log file he had 3000 filedescriptors, problems with the messages that there were no filedescriptors more available. After I increase also the NUMSP value the problems were gone.
One thing left is the number of tcp-connections the system can handle. Default is 256, but I increase that as well because of the number of clients we have.
32-bit processes on AIX and later are restricted by default to a maximum
of 11 shared memory segments. This restriction can be removed on AIX 4.2.1
and later by setting the environment variable EXTSHM=ON in the script or
shell which starts squid.
32-bit processes cannot use more than 256MB of stack and data in the default memory model. To force the loader to use large address space for squid, either:
LDR_CNTRL environment variable,
eg LDR_CNTRL="MAXDATA=0x80000000"; or-bmaxdata:0x80000000; orSquid has the ability to rewrite requested URLs. Implemented as an external process (similar to a dnsserver), Squid can be configured to pass every incoming URL through a redirector process that returns either a new URL, or a blank line to indicate no change.
The redirector program is NOT a standard part of the Squid package. However, some examples are provided below, and in the "contrib/" directory of the source distribution. Since everyone has different needs, it is up to the individual administrators to write their own implementation.
A redirector allows the administrator to control the locations to which his users goto. Using this in conjunction with interception proxies allows simple but effective porn control.
The redirector program must read URLs (one per line) on standard input, and write rewritten URLs or blank lines on standard output. Note that the redirector program can not use buffered I/O. Squid writes additional information after the URL which a redirector can use to make a decision. The input line consists of four fields:
URL ip-address/fqdn ident method
A simple very fast redirector called SQUIRM is a good place to start, it uses the regex lib to allow pattern matching.
Also see jesred.
The following Perl script may also be used as a template for writing your own redirector:
#!/usr/local/bin/perl
$|=1;
while (<>) {
s@http://fromhost.com@http://tohost.org@;
print;
}
Normally, the redirector feature is used to rewrite requested URLs. Squid then transparently requests the new URL. However, in some situations, it may be desirable to return an HTTP "301" or "302" redirect message to the client. This is now possible with Squid version 1.1.19.
Simply modify your redirector program to prepend either "301:" or "302:" before the new URL. For example, the following script might be used to direct external clients to a secure Web server for internal documents:
#!/usr/local/bin/perl
$|=1;
while (<>) {
@X = split;
$url = $X[0];
if ($url =~ /^http:\/\/internal\.foo\.com/) {
$url =~ s/^http/https/;
$url =~ s/internal/secure/;
print "302:$url\n";
} else {
print "$url\n";
}
}
Please see sections 10.3.2 and 10.3.3 of RFC 2068 for an explanation of the 301 and 302 HTTP reply codes.
A redirector process must exit (stop running) only when its stdin is closed. If you see the ``All redirectories have exited'' message, it probably means your redirector program has a bug. Maybe it runs out of memory or has memory access errors. You may want to test your redirector program outside of squid with a big input list, taken from your access.log perhaps. Also, check for coredump files from the redirector program.
I added a redirctor consisting of
#! /bin/sh /usr/bin/tee /tmp/squid.logand many of the redirector requests don't have a username in the ident field.
Squid does not delay a request to wait for an ident lookup, unless you use the ident ACLs. Thus, it is very likely that the ident was not available at the time of calling the redirector, but became available by the time the request is complete and logged to access.log.
If you want to block requests waiting for ident lookup, try something like this:
acl foo ident REQUIRED http_access allow foo
Cache Digest FAQs compiled by Niall Doherty.
A Cache Digest is a summary of the contents of an Internet Object Caching Server. It contains, in a compact (i.e. compressed) format, an indication of whether or not particular URLs are in the cache.
A "lossy" technique is used for compression, which means that very high compression factors can be achieved at the expense of not having 100% correct information.
Cache servers periodically exchange their digests with each other.
When a request for an object (URL) is received from a client a cache can use digests from its peers to find out which of its peers (if any) have that object. The cache can then request the object from the closest peer (Squid uses the NetDB database to determine this).
Note that Squid will only make digest queries in those digests that are enabled. It will disable a peers digest IFF it cannot fetch a valid digest for that peer. It will enable that peers digest again when a valid one is fetched.
The checks in the digest are very fast and they eliminate the need for per-request queries to peers. Hence:
Note that the use of Cache Digests (for querying the cache contents of peers) and the generation of a Cache Digest (for retrieval by peers) are independent. So, it is possible for a cache to make a digest available for peers, and not use the functionality itself and vice versa.
Cache Digests are based on Bloom Filters - they are a method for representing a set of keys with lookup capabilities; where lookup means "is the key in the filter or not?".
In building a cache digest:
To add a key the value of each hash function for that key is calculated. So, if the key was denoted by a, then h1(a), h2(a), ..., hk(a) are calculated.
The value of each hash function for that key represents an index into the array and the corresponding bits are set to 1. So, a digest with 6 hash functions would have 6 bits to be set to 1 for each key added.
Note that the addition of a number of different keys could cause one particular bit to be set to 1 multiple times.
To query for the existence of a key the indices into the array are calculated from the hash functions as above.
Note the term likely. It is possible that a collision in the digest can occur, whereby the digest incorrectly indicates a key is present. This is the price paid for the compact representation. While the probability of a collision can never be reduced to zero it can be controlled. Larger values for the ratio of the digest size to the number of entries added lower the probability. The number of hash functions chosen also influence the probability.
To delete a key, it is not possible to simply set the associated bits to 0 since any one of those bits could have been set to 1 by the addition of a different key!
Therefore, to support deletions a counter is required for each bit position in the array. The procedures to follow would be:
Upon initialisation, the capacity is set to the number of objects that can be (are) stored in the cache. Note that there are upper and lower limits here.
An arbitrary constant, bits_per_entry (currently set to 5), is used to calculate the size of the array using the following formula:
number of bits in array = capacity * bits_per_entry + 7
The size of the digest, in bytes, is therefore:
digest size = int (number of bits in array / 8)
When a digest rebuild occurs, the change in the cache size (capacity) is measured. If the capacity has changed by a large enough amount (10%) then the digest array is freed and reallocated memory, otherwise the same digest is re-used.
The protocol design allows for a variable number of hash functions (k). However, Squid employs a very efficient method using a fixed number - four.
Rather than computing a number of independent hash functions over a URL Squid uses a 128-bit MD5 hash of the key (actually a combination of the URL and the HTTP retrieval method) and then splits this into four equal chunks.
Each chunk, modulo the digest size (m), is used as the value for one of the hash functions - i.e. an index into the bit array.
Note: As Squid retrieves objects and stores them in its cache on disk, it adds them to the in-RAM index using a lookup key which is an MD5 hash - the very one discussed above. This means that the values for the Cache Digest hash functions are already available and consequently the operations are extremely efficient!
Obviously, modifying the code to support a variable number of hash functions would prove a little more difficult and would most likely reduce efficiency.
Every object referenced in the index in RAM is checked to see if it is suitable for addition to the digest.
A number of objects are not suitable, e.g. those that are private, not cachable, negatively cached etc. and are skipped immediately.
A freshness test is next made in an attempt to guess if the object will expire soon, since if it does, it is not worthwhile adding it to the digest. The object is checked against the refresh patterns for staleness...
Since Squid stores references to objects in its index using the MD5 key discussed earlier there is no URL actually available for each object - which means that the pattern used will fall back to the default pattern, ".". This is an unfortunate state of affairs, but little can be done about it. A cd_refresh_pattern option will be added to the configuration file soon which will at least make the confusion a little clearer :-)
Note that it is best to be conservative with your refresh pattern for the Cache Digest, i.e. do not add objects if they might become stale soon. This will reduce the number of False Hits.
Squid does not support deletions from the digest. Because of this the digest must, periodically, be rebuilt from scratch to erase stale bits and prevent digest pollution.
A more sophisticated option is to use diffs or deltas. These would be created by building a new digest and comparing with the current/old one. They would essentially consist of aggregated deletions and additions since the previous digest.
Since less bandwidth should be required using these it would be possible to have more frequent updates (and hence, more accurate information).
Costs:
The local digest is built:
While the [new] digest is being built in RAM the old version (stored on disk) is still valid, and will be returned to any peer requesting it. When the digest has completed building it is then swapped out to disk, overwriting the old version.
The rebuild is CPU intensive, but not overly so. Since Squid is programmed using an event-handling model, the approach taken is to split the digest building task into chunks (i.e. chunks of entries to add) and to register each chunk as an event. If CPU load is overly high, it is possible to extend the build period - as long as it is finished before the next rebuild is due!
It may prove more efficient to implement the digest building as a separate process/thread in the future...
Cache Digests are fetched from peers using the standard HTTP protocol (note that a pull rather than push technique is used).
After the first access to a peer, a peerDigestValidate event is queued (this event decides if it is time to fetch a new version of a digest from a peer). The queuing delay depends on the number of peers already queued for validation - so that all digests from different peers are not fetched simultaneously.
A peer answering a request for its digest will specify an expiry time for that digest by using the HTTP Expires header. The requesting cache thus knows when it should request a fresh copy of that peers digest.
Note: requesting caches use an If-Modified-Since request in case the peer has not rebuilt its digest for some reason since the last time it was fetched.
Since the local digest is generated purely for the benefit of its neighbours keeping it in RAM is not strictly required. However, it was decided to keep the local digest in RAM partly because of the following:
When the digest is built in RAM, it is then swapped out to disk, where it is stored as a "normal" cache item - which is how peers request it.
When a query from a client arrives, fast lookups are required to decide if a request should be made to a neighbour cache. It it therefore required to keep all peer digests in RAM.
Peer digests are also stored on disk for the following reasons:
Cache Digest statistics can be seen from the Cache Manager or through the squidclient utility. The following examples show how to use the squidclient utility to request the list of possible operations from the localhost, local digest statistics from the localhost, refresh statistics from the localhost and local digest statistics from another cache, respectively.
squidclient mgr:menu squidclient mgr:store_digest squidclient mgr:refresh squidclient -h peer mgr:store_digest
The available statistics provide a lot of useful debugging information. The refresh statistics include a section for Cache Digests which explains why items were added (or not) to the digest.
The following example shows local digest statistics for a 16GB cache in a corporate intranet environment (may be a useful reference for the discussion below).
store digest: size: 768000 bytes entries: count: 588327 capacity: 1228800 util: 48% deletion attempts: 0 bits: per entry: 5 on: 1953311 capacity: 6144000 util: 32% bit-seq: count: 2664350 avg.len: 2.31 added: 588327 rejected: 528703 ( 47.33 %) del-ed: 0 collisions: on add: 0.23 % on rej: 0.23 %
entries:capacity is a measure of how many items "are likely" to be added to the digest. It represents the number of items that were in the local cache at the start of digest creation - however, upper and lower limits currently apply. This value is multiplied by bits: per entry (an arbitrary constant) to give bits:capacity, which is the size of the cache digest in bits. Dividing this by 8 will give store digest: size which is the size in bytes.
The number of items represented in the digest is given by entries:count. This should be equal to added minus deletion attempts.
Since (currently) no modifications are made to the digest after the initial build (no additions are made and deletions are not supported) deletion attempts will always be 0 and entries:count should simply be equal to added.
entries:util is not really a significant statistic. At most it gives a measure of how many of the items in the store were deemed suitable for entry into the cache compared to how many were "prepared" for.
rej shows how many objects were rejected. Objects will not be added for a number of reasons, the most common being refresh pattern settings. Remember that (currently) the default refresh pattern will be used for checking for entry here and also note that changing this pattern can significantly affect the number of items added to the digest! Too relaxed and False Hits increase, too strict and False Misses increase. Remember also that at time of validation (on the peer) the "real" refresh pattern will be used - so it is wise to keep the default refresh pattern conservative.
bits: on indicates the number of bits in the digest that are set to 1. bits: util gives this figure as a percentage of the total number of bits in the digest. As we saw earlier, a figure of 50% represents the optimal trade-off. Values too high (say > 75%) would cause a larger number of collisions, and hence False Hits, while lower values mean the digest is under-utilised (using unnecessary RAM). Note that low values are normal for caches that are starting to fill up.
A bit sequence is an uninterrupted sequence of bits with the same value. bit-seq: avg.len gives some insight into the quality of the hash functions. Long values indicate problem, even if bits:util is 50% (> 3 = suspicious, > 10 = very suspicious).
A False Hit occurs when a cache believes a peer has an object and asks the peer for it but the peer is not able to satisfy the request.
Expiring or stale objects on the peer are frequent causes of False Hits. At the time of the query actual refresh patterns are used on the peer and stale entries are marked for revalidation. However, revalidation is prohibited unless the peer is behaving as a parent, or miss_access is enabled. Thus, clients can receive error messages instead of revalidated objects!
The frequency of False Hits can be reduced but never eliminated completely, therefore there must be a robust way of handling them when they occur. The philosophy behind the design of Squid is to use lightweight techniques and optimise for the common case and robustly handle the unusual case (False Hits).
Squid will soon support the HTTP only-if-cached header. Requests for objects made to a peer will use this header and if the objects are not available, the peer can reply appropriately allowing Squid to recognise the situation. The following describes what Squid is aiming towards:
If you wish to use Cache Digests (available in Squid version 2) you need to add a configure option, so that the relevant code is compiled in:
./configure --enable-cache-digests ...
If a request is forwarded to a neighbour due a HIT in that neighbour's Cache Digest the hierarchy (9th) field of the access.log file for the local cache will look like CACHE_DIGEST_HIT/neighbour. The Log Tag (4th field) should obviously show a MISS.
On the peer cache the request should appear as a normal HTTP request from the first cache.
The easiest situation to analyse is when two caches (say A and B) are involved neither of which uses the other as a parent. In this case, a False Hit would show up as a CACHE_DIGEST_HIT on A and NOT as a TCP_HIT on B (or vice versa). If B does not fetch the object for A then the hierarchy field will look like NONE/- (and A will have received an Access Denied or Forbidden message). This will happen if the object is not "available" on B and B does not have miss_access enabled for A (or is not acting as a parent for A).
Assume A requests a URL from B and receives a False Hit
squidclient -m PURGE 'URL'
squidclient 'URL'
The HTTP headers of the request are available. Two header types are of particular interest:
The X-Cache and X-Cache-Lookup headers from A should both show MISS.
If A requests the object from B (which it will if the digest lookup indicates B has it - assuming B is closest peer of course :-) then there will be another set of these headers from B.
If the X-Cache header from B shows a MISS a False Hit has occurred. This means that A thought B had an object but B tells A it does not have it available for retrieval. The reason why it is not available for retrieval is indicated by the X-Cache-Lookup header. If:
If there is something else you need to check you can always look at the source code. The main Cache Digest functionality is organised as follows:
Note that in the source the term Store Digest refers to the digest created locally. The Cache Digest code is fairly self-explanatory (once you understand how Cache Digests work):
COMING SOON!
There is now, thanks to Martin Hamilton and Alex Rousskov.
Cache Digests, as implemented in Squid 2.1.PATCH2, are described in cache-digest-v5.txt.
You'll notice the format is similar to an Internet Draft. We decided not to submit this document as a draft because Cache Digests will likely undergo some important changes before we want to try to make it a standard.
Note: The information here is current for version 2.2.
Squid already has code to spread the digest updates. The algorithm is currently controlled by a few hard-coded constants in peer_digest.c. For example, GlobDigestReqMinGap variable determines the minimum interval between two requests for a digest. You may want to try to increase the value of GlobDigestReqMinGap from 60 seconds to whatever you feel comfortable with (but it should be smaller than hour/number_of_peers, of course).
Note that whatever you do, you still need to give Squid enough time and bandwidth to fetch all the digests. Depending on your environment, that bandwidth may be more or less than an ICP would require. Upcoming digest deltas (x10 smaller than the digests themselves) may be the only way to solve the ``big scale'' problem.
How can I make my users' browsers use my cache without configuring the browsers for proxying?
First, it is critical to read the full comments in the squid.conf file! That is the only authoritative source for configuration information. However, the following instructions are correct as of this writing (July 1999.)
Getting interception caching to work requires four distinct steps:
http_port 8080 httpd_accel_host virtual httpd_accel_port 80 httpd_accel_with_proxy on httpd_accel_uses_host_header on
Notes:
NOTE: You don't need to use IP Filter on FreeBSD. Use the built-in ipfw feature instead. See the FreeBSD subsection below.
First, get and install the IP Filter package.
Put these lines in /etc/ipnat.rules:
# Redirect direct web traffic to local web server. rdr de0 1.2.3.4/32 port 80 -> 1.2.3.4 port 80 tcp # Redirect everything else to squid on port 8080 rdr de0 0.0.0.0/0 port 80 -> 1.2.3.4 port 8080 tcp
Modify your startup scripts to enable ipnat. For example, on FreeBSD it looks something like this:
/sbin/modload /lkm/if_ipl.o /sbin/ipnat -f /etc/ipnat.rules chgrp nobody /dev/ipnat chmod 644 /dev/ipnat
Squid-2 (after version beta25) has IP filter support built in. Simple enable it when you run configure:
./configure --enable-ipf-transparentAdd these lines to your squid.conf file:
http_port 8080 httpd_accel_host virtual httpd_accel_port 80 httpd_accel_with_proxy on httpd_accel_uses_host_header onNote, you don't have to use port 8080, but it must match whatever you used in the /etc/ipnat.rules file.
Patches for Squid-1.X are available from Quinton Dolan's Squid page. Add these lines to squid.conf:
http_port 8080 httpd_accel virtual 80 httpd_accel_with_proxy on httpd_accel_uses_host_header on
Thanks to Quinton Dolan.
Note: Interception proxying does NOT work with Linux 2.0.30! Linux 2.0.29 is known to work well. If you're using a more recent kernel, like 2.2.X, then you should probably use an ipchains configuration, as described below.
Warning: this technique has some shortcomings.
If you can live with the side-effects, go ahead and compile your kernel with firewalling and redirection support. Here are the important parameters from /usr/src/linux/.config:
# # Code maturity level options # CONFIG_EXPERIMENTAL=y # # Networking options # CONFIG_FIREWALL=y # CONFIG_NET_ALIAS is not set CONFIG_INET=y CONFIG_IP_FORWARD=y # CONFIG_IP_MULTICAST is not set CONFIG_IP_FIREWALL=y # CONFIG_IP_FIREWALL_VERBOSE is not set CONFIG_IP_MASQUERADE=y CONFIG_IP_TRANSPARENT_PROXY=y CONFIG_IP_ALWAYS_DEFRAG=y # CONFIG_IP_ACCT is not set CONFIG_IP_ROUTER=y
You may also need to enable IP Forwarding. One way to do it is to add this line to your startup scripts:
echo 1 > /proc/sys/net/ipv4/ip_forward
Go to the Linux IP Firewall and Accounting page, obtain the source distribution to ipfwadm and install it. Older versions of ipfwadm may not work. You might need at least version 2.3.0. You'll use ipfwadm to setup the redirection rules. I added this rule to the script that runs from /etc/rc.d/rc.inet1 (Slackware) which sets up the interfaces at boot-time. The redirection should be done before any other Input-accept rule. To really make sure it worked I disabled the forwarding (masquerading) I normally do.
/etc/rc.d/rc.firewall:
#!/bin/sh
# rc.firewall Linux kernel firewalling rules
FW=/sbin/ipfwadm
# Flush rules, for testing purposes
for i in I O F # A # If we enabled accounting too
do
${FW} -$i -f
done
# Default policies:
${FW} -I -p rej # Incoming policy: reject (quick error)
${FW} -O -p acc # Output policy: accept
${FW} -F -p den # Forwarding policy: deny
# Input Rules:
# Loopback-interface (local access, eg, to local nameserver):
${FW} -I -a acc -S localhost/32 -D localhost/32
# Local Ethernet-interface:
# Redirect to Squid proxy server:
${FW} -I -a acc -P tcp -D default/0 80 -r 8080
# Accept packets from local network:
${FW} -I -a acc -P all -S localnet/8 -D default/0 -W eth0
# Only required for other types of traffic (FTP, Telnet):
# Forward localnet with masquerading (udp and tcp, no icmp!):
${FW} -F -a m -P tcp -S localnet/8 -D default/0
${FW} -F -a m -P udp -S localnet/8 -D default/0
Here all traffic from the local LAN with any destination gets redirected to the local port 8080. Rules can be viewed like this:
IP firewall input rules, default policy: reject type prot source destination ports acc all 127.0.0.1 127.0.0.1 n/a acc/r tcp 10.0.0.0/8 0.0.0.0/0 * -> 80 => 8080 acc all 10.0.0.0/8 0.0.0.0/0 n/a acc tcp 0.0.0.0/0 0.0.0.0/0 * -> *
I did some testing on Windows 95 with both Microsoft Internet Explorer 3.01 and Netscape Communicator pre-release and it worked with both browsers with the proxy-settings disabled.
At one time squid seemed to get in a loop when I pointed the browser to the local port 80. But this could be avoided by adding a reject rule for client to this address:
${FW} -I -a rej -P tcp -S localnet/8 -D hostname/32 80
IP firewall input rules, default policy: reject
type prot source destination ports
acc all 127.0.0.1 127.0.0.1 n/a
rej tcp 10.0.0.0/8 10.0.0.1 * -> 80
acc/r tcp 10.0.0.0/8 0.0.0.0/0 * -> 80 => 8080
acc all 10.0.0.0/8 0.0.0.0/0 n/a
acc tcp 0.0.0.0/0 0.0.0.0/0 * -> *
NOTE on resolving names: Instead of just passing the URLs to the proxy server, the browser itself has to resolve the URLs. Make sure the workstations are setup to query a local nameserver, to minimize outgoing traffic.
If you're already running a nameserver at the firewall or proxy server (which is a good idea anyway IMHO) let the workstations use this nameserver.
Additional notes from Richard Ayres
I'm using such a setup. The only issues so far have been that:
- It's fairly useless to use my service providers parent caches (cache-?.www.demon.net) because by proxying squid only sees IP addresses, not host names and demon aren't generally asked for IP addresses by other users;
- Linux kernel 2.0.30 is a no-no as interception proxying is broken (I use 2.0.29);
- Client browsers must do host name lookups themselves, as they don't know they're using a proxy;
- The Microsoft Network won't authorize its users through a proxy, so I have to specifically *not* redirect those packets (my company is a MSN content provider).
Aside from this, I get a 30-40% hit rate on a 50MB cache for 30-40 users and am quite pleased with the results.
See also Daniel Kiracofe's page.
by Martin Lyons
You need to configure your kernel for ipchains. Configuring Linux kernels is beyond the scope of this FAQ. One way to do it is:
# cd /usr/src/linux # make menuconfig
The following shows important kernel features to include:
[*] Network firewalls [ ] Socket Filtering [*] Unix domain sockets [*] TCP/IP networking [ ] IP: multicasting [ ] IP: advanced router [ ] IP: kernel level autoconfiguration [*] IP: firewalling [ ] IP: firewall packet netlink device [*] IP: always defragment (required for masquerading) [*] IP: transparent proxy support
You must include the IP: always defragment, otherwise it prevents you from using the REDIRECT chain.
You can use this script as a template for your own rc.firewall to configure ipchains:
#!/bin/sh
# rc.firewall Linux kernel firewalling rules
# Leon Brooks (leon at brooks dot fdns dot net)
FW=/sbin/ipchains
ADD="$FW -A"
# Flush rules, for testing purposes
for i in I O F # A # If we enabled accounting too
do
${FW} -F $i
done
# Default policies:
${FW} -P input REJECT # Incoming policy: reject (quick error)
${FW} -P output ACCEPT # Output policy: accept
${FW} -P forward DENY # Forwarding policy: deny
# Input Rules:
# Loopback-interface (local access, eg, to local nameserver):
${ADD} input -j ACCEPT -s localhost/32 -d localhost/32
# Local Ethernet-interface:
# Redirect to Squid proxy server:
${ADD} input -p tcp -d 0/0 80 -j REDIRECT 8080
# Accept packets from local network:
${ADD} input -j ACCEPT -s localnet/8 -d 0/0 -i eth0
# Only required for other types of traffic (FTP, Telnet):
# Forward localnet with masquerading (udp and tcp, no icmp!):
${ADD} forward -j MASQ -p tcp -s localnet/8 -d 0/0
${ADD} forward -j MASQ -P udp -s localnet/8 -d 0/0
Also, Andrew Shipton notes that with 2.0.x kernels you don't need to enable packet forwarding, but with the 2.1.x and 2.2.x kernels using ipchains you do. Packet forwarding is enabled with the following command:
echo 1 > /proc/sys/net/ipv4/ip_forward
NOTE: this information comes from Daniel Kiracofe's Transparent Proxy with Squid mini-HOWTO.
To support netfilter transparent interception on Linux 2.4 Squid must be compiled with the --enable-linux-netfilter option.
To enable netwfilter support you may need to build a new kernel. Be sure to enable all of these options:
You must say NO to ``Fast switching''
After building the kernel, install it and reboot.
You may need to enable packet forwarding (e.g. in your startup scripts):
echo 1 > /proc/sys/net/ipv4/ip_forward
Use the iptables command to make your kernel intercept HTTP connections and send them to Squid:
iptables -t nat -A PREROUTING -i eth0 -p tcp --dport 80 -j REDIRECT --to-port 3128
This works with at least IOS 11.1 and later I guess. Possibly earlier, as I'm no CISCO expert I can't say for sure. If your router is doing anything more complicated that shuffling packets between an ethernet interface and either a serial port or BRI port, then you should work through if this will work for you.
First define a route map with a name of proxy-redirect (name doesn't matter) and specify the next hop to be the machine Squid runs on.
! route-map proxy-redirect permit 10 match ip address 110 set ip next-hop 203.24.133.2 !Define an access list to trap HTTP requests. The second line allows the Squid host direct access so an routing loop is not formed. By carefully writing your access list as show below, common cases are found quickly and this can greatly reduce the load on your router's processor.
! access-list 110 deny tcp any any neq www access-list 110 deny tcp host 203.24.133.2 any access-list 110 permit tcp any any !Apply the route map to the ethernet interface.
! interface Ethernet0 ip policy route-map proxy-redirect !
Bruce Morgan notes that there is a Cisco bug relating to interception proxying using IP policy route maps, that causes NFS and other applications to break. Apparently there are two bug reports raised in Cisco, but they are not available for public dissemination.
The problem occurs with o/s packets with more than 1472 data bytes. If you try to ping a host with more than 1472 data bytes across a Cisco interface with the access-lists and ip policy route map, the icmp request will fail. The packet will be fragmented, and the first fragment is checked against the access-list and rejected - it goes the "normal path" as it is an icmp packet - however when the second fragment is checked against the access-list it is accepted (it isn't regarded as an icmp packet), and goes to the action determined by the policy route map!
John notes that you may be able to get around this bug by carefully writing your access lists. If the last/default rule is to permit then this bug would be a problem, but if the last/default rule was to deny then it won't be a problem. I guess fragments, other than the first, don't have the information available to properly policy route them. Normally TCP packets should not be fragmented, at least my network runs an MTU of 1500 everywhere to avoid fragmentation. So this would affect UDP and ICMP traffic only.
Basically, you will have to pick between living with the bug or better performance. This set has better performance, but suffers from the bug:
access-list 110 deny tcp any any neq www access-list 110 deny tcp host 10.1.2.3 any access-list 110 permit tcp any anyConversely, this set has worse performance, but works for all protocols:
access-list 110 deny tcp host 10.1.2.3 any access-list 110 permit tcp any any eq www access-list 110 deny tcp any any
Just for kicks, here's an email message posted to squid-users on how to make interception proxying work with a Cisco router and Squid running on Linux.
by Brian Feeny
Here is how I have Interception proxying working for me, in an environment where my router is a Cisco 2501 running IOS 11.1, and Squid machine is running Linux 2.0.33.
Many thanks to the following individuals and the squid-users list for helping me get redirection and interception proxying working on my Cisco/Linux box.
First, here is what I added to my Cisco, which is running IOS 11.1. In IOS 11.1 the route-map command is "process switched" as opposed to the faster "fast-switched" route-map which is found in IOS 11.2 and later. You may wish to be running IOS 11.2. I am running 11.1, and have had no problems with my current load of about 150 simultaneous connections to squid.:
! interface Ethernet0 description To Office Ethernet ip address 208.206.76.1 255.255.255.0 no ip directed-broadcast no ip mroute-cache ip policy route-map proxy-redir ! access-list 110 deny tcp host 208.206.76.44 any eq www access-list 110 permit tcp any any eq www route-map proxy-redir permit 10 match ip address 110 set ip next-hop 208.206.76.44
So basically from above you can see I added the "route-map" declaration, and an access-list, and then turned the route-map on under int e0 "ip policy route-map proxy-redir"
ok, so the Cisco is taken care of at this point. The host above: 208.206.76.44, is the ip number of my squid host.
My squid box runs Linux, so I had to do the following on it:
my kernel (2.0.33) config looks like this:
# # Networking options # CONFIG_FIREWALL=y # CONFIG_NET_ALIAS is not set CONFIG_INET=y CONFIG_IP_FORWARD=y CONFIG_IP_MULTICAST=y CONFIG_SYN_COOKIES=y # CONFIG_RST_COOKIES is not set CONFIG_IP_FIREWALL=y # CONFIG_IP_FIREWALL_VERBOSE is not set CONFIG_IP_MASQUERADE=y # CONFIG_IP_MASQUERADE_IPAUTOFW is not set CONFIG_IP_MASQUERADE_ICMP=y CONFIG_IP_TRANSPARENT_PROXY=y CONFIG_IP_ALWAYS_DEFRAG=y # CONFIG_IP_ACCT is not set CONFIG_IP_ROUTER=y
You will need Firewalling and Transparent Proxy turned on at a minimum.
Then some ipfwadm stuff:
# Accept all on loopback ipfwadm -I -a accept -W lo # Accept my own IP, to prevent loops (repeat for each interface/alias) ipfwadm -I -a accept -P tcp -D 208.206.76.44 80 # Send all traffic destined to port 80 to Squid on port 3128 ipfwadm -I -a accept -P tcp -D 0/0 80 -r 3128
it accepts packets on port 80 (redirected from the Cisco), and redirects them to 3128 which is the port my squid process is sitting on. I put all this in /etc/rc.d/rc.local
I am using v1.1.20 of Squid with Henrik's patch installed. You will want to install this patch if using a setup similar to mine.
by Duane Wessels
I set out yesterday to make interception caching work with Squid and FreeBSD. It was, uh, fun.
It was relatively easy to configure a cisco to divert port 80 packets to my FreeBSD box. Configuration goes something like this:
access-list 110 deny tcp host 10.0.3.22 any eq www access-list 110 permit tcp any any eq www route-map proxy-redirect permit 10 match ip address 110 set ip next-hop 10.0.3.22 int eth2/0 ip policy route-map proxy-redirectHere, 10.0.3.22 is the IP address of the FreeBSD cache machine.
Once I have packets going to the FreeBSD box, I need to get the kernel to deliver them to Squid. I started on FreeBSD-2.2.7, and then downloaded IPFilter. This was a dead end for me. The IPFilter distribution includes patches to the FreeBSD kernel sources, but many of these had conflicts. Then I noticed that the IPFilter page says ``It comes as a part of [FreeBSD-2.2 and later].'' Fair enough. Unfortunately, you can't hijack connections with the FreeBSD-2.2.X IPFIREWALL code (ipfw), and you can't (or at least I couldn't) do it with natd either.
FreeBSD-3.0 has much better support for connection hijacking, so I suggest you start with that. You need to build a kernel with the following options:
options IPFIREWALL options IPFIREWALL_FORWARD
Next, its time to configure the IP firewall rules with ipfw. By default, there are no "allow" rules and all packets are denied. I added these commands to /etc/rc.local just to be able to use the machine on my network:
ipfw add 60000 allow all from any to anyBut we're still not hijacking connections. To accomplish that, add these rules:
ipfw add 49 allow tcp from 10.0.3.22 to any ipfw add 50 fwd 127.0.0.1 tcp from any to any 80The second line (rule 50) is the one which hijacks the connection. The first line makes sure we never hit rule 50 for traffic originated by the local machine. This prevents forwarding loops.
Note that I am not changing the port number here. That is, port 80 packets are simply diverted to Squid on port 80. My Squid configuration is:
http_port 80 httpd_accel_host virtual httpd_accel_port 80 httpd_accel_with_proxy on httpd_accel_uses_host_header on
If you don't want Squid to listen on port 80 (because that requires root privileges) then you can use another port. In that case your ipfw redirect rule looks like:
ipfw add 50 fwd 127.0.0.1,3128 tcp from any to any 80and the squid.conf lines are:
http_port 3128 httpd_accel_host virtual httpd_accel_port 80 httpd_accel_with_proxy on httpd_accel_uses_host_header on
This is to do with configuring interception proxy for an ACC Tigris digital access server (like a CISCO 5200/5300 or an Ascend MAX 4000). I've found that doing this in the NAS reduces traffic on the LAN and reduces processing load on the CISCO. The Tigris has ample CPU for filtering.
Step 1 is to create filters that allow local traffic to pass. Add as many as needed for all of your address ranges.
ADD PROFILE IP FILTER ENTRY local1 INPUT 10.0.3.0 255.255.255.0 0.0.0.0 0.0.0.0 NORMAL ADD PROFILE IP FILTER ENTRY local2 INPUT 10.0.4.0 255.255.255.0 0.0.0.0 0.0.0.0 NORMAL
Step 2 is to create a filter to trap port 80 traffic.
ADD PROFILE IP FILTER ENTRY http INPUT 0.0.0.0 0.0.0.0 0.0.0.0 0.0.0.0 = 0x6 D= 80 NORMAL
Step 3 is to set the "APPLICATION_ID" on port 80 traffic to 80. This causes all packets matching this filter to have ID 80 instead of the default ID of 0.
SET PROFILE IP FILTER APPLICATION_ID http 80
Step 4 is to create a special route that is used for packets with "APPLICATION_ID" set to 80. The routing engine uses the ID to select which routes to use.
ADD IP ROUTE ENTRY 0.0.0.0 0.0.0.0 PROXY-IP 1 SET IP ROUTE APPLICATION_ID 0.0.0.0 0.0.0.0 PROXY-IP 80
Step 5 is to bind everything to a filter ID called transproxy. List all local filters first and the http one last.
ADD PROFILE ENTRY transproxy local1 local2 http
With this in place use your RADIUS server to send back the ``Framed-Filter-Id = transproxy'' key/value pair to the NAS.
You can check if the filter is being assigned to logins with the following command:
display profile port table
by Brian Feeny.
First, configure Squid for interception caching as detailed at the beginning of this section.
Next, configure the Foundry layer 4 switch to redirect traffic to your Squid box or boxes. By default, the Foundry redirects to port 80 of your squid box. This can be changed to a different port if needed, but won't be covered here.
In addition, the switch does a "health check" of the port to make sure your squid is answering. If you squid does not answer, the switch defaults to sending traffic directly thru instead of redirecting it. When the Squid comes back up, it begins redirecting once again.
This example assumes you have two squid caches:
squid1.foo.com 192.168.1.10 squid2.foo.com 192.168.1.11
We will assume you have various workstations, customers, etc, plugged into the switch for which you want them to be intercepted and sent to Squid. The squid caches themselves should be plugged into the switch as well. Only the interface that the router is connected to is important. Where you put the squid caches or other connections does not matter.
This example assumes your router is plugged into interface 17 of the switch. If not, adjust the following commands accordingly.
telnet@ServerIron#conf t
telnet@ServerIron(config)# server cache-name squid1 192.168.1.10 telnet@ServerIron(config)# server cache-name squid2 192.168.1.11
telnet@ServerIron(config)#server cache-group 1 telnet@ServerIron(config-tc-1)#cache-name squid1 telnet@ServerIron(config-tc-1)#cache-name squid2
telnet@ServerIron(config)# ip policy 1 cache tcp http local
telnet@ServerIron(config)#int e 17 telnet@ServerIron(config-if-17)# ip-policy 1
Since all outbound traffic to the Internet goes out interface 17 (the router), and interface 17 has the caching policy applied to it, HTTP traffic is going to be intercepted and redirected to the caches you have configured.
The default port to redirect to can be changed. The load balancing algorithm used can be changed (Least Used, Round Robin, etc). Ports can be exempted from caching if needed. Access Lists can be applied so that only certain source IP Addresses are redirected, etc. This information was left out of this document since this was just a quick howto that would apply for most people, not meant to be a comprehensive manual of how to configure a Foundry switch. I can however revise this with any information necessary if people feel it should be included.
By Dave Wintrip, dave at purevanity dot net, June 3, 2004.
I have verified this configuration as working on a Cabletron SmartSwitchRouter 2000, and it should work on any layer-4 aware Cabletron or Entrasys product.
You must first configure Squid to enable interception caching, outlined earlier.
Next, make sure that you have connectivity from the layer-4 device to your squid box, and that squid is correctly configured to intercept port 80 requests thrown it's way.
I generally create two sets of redirect ACLs, one for cache, and one for bypassing the cache. This method of interception is very similar to Cisco's route-map.
Log into the device, and enter enable mode, as well as configure mode.
ssr> en Password: ssr# conf ssr(conf)#
I generally create two sets of redirect ACLs, one for specifying who to cache, and one for destination addresses that need to bypass the cache. This method of interception is very similar to Cisco's route-map in this way. The ACL cache-skip is a list of destination addresses that we do not want to transparently redirect to squid.
ssr(conf)# acl cache-skip permit tcp any 192.168.1.100/255.255.255.255 any http
The ACL cache-allow is a list of source addresses that will be redirected to Squid.
ssr(conf)# acl cache-allow permit tcp 10.0.22.0/255.255.255.0 any any http
Save your new ACLs to the running configuration.
ssr(conf)# save a
Next, we need to create the ip-policies that will work to perform the redirection. Please note that 10.0.23.2 is my Squid server, and that 10.0.24.1 is my standard default next hop. By pushing the cache-skip ACL to the default gateway, the web request is sent out as if the squid box was not present. This could just as easily be done using the squid configuration, but I would rather Squid not touch the data if it has no reason to.
ssr(conf)# ip-policy cache-allow permit acl cache-allow next-hop-list 10.0.23.2 action policy-only ssr(conf)# ip-policy cache-skip permit acl cache-skip next-hop-list 10.0.24.1 action policy-only
Apply these new policies into the active configuration.
ssr(conf)# save a
We now need to apply the ip-policies to interfaces we want to cache requests from. Assuming that localnet-gw is the interface name to the network we want to cache requests from, we first apply the cache-skip ACL to intercept requests on our do-not-cache list, and forward them out the default gateway. We then apply the cache-allow ACL to the same interface to redirect all other requests to the cache server.
ssr(conf)# ip-policy cache-skip apply interface localnet-gw ssr(conf)# ip-policy cache-allow apply interface localnet-gw
We now need to apply, and permanently save our changes. Nothing we have done before this point would effect anything without adding the ip-policy applications into the active configuration, so lets try it.
ssr(conf)# save a ssr(conf)# save s
Provided your Squid box is correct configured, you should now be able to surf, and be transparently cached if you are using the localnet-gw address as your gateway.
Some Cabletron/Entrasys products include another method of applying a web cache, but details on configuring that is not covered in this document, however is it fairly straight forward.
Also note, that if your Squid box is plugged directly into a port on your layer-4 switch, and that port is part of its own VLAN, and its own subnet, if that port were to change states to down, or the address becomes uncontactable, then the switch will automatically bypass the ip-policies and forward your web request though the normal means. This is handy, might I add.
I think almost everyone who have tried to build a interception proxy setup have been bitten by this one.
Measures you can take:
Fyodor has tracked down the cause of unusual ``connection reset by peer'' messages when using Cisco policy routing to hijack HTTP requests.
When the network link between router and the cache goes down for just a moment, the packets that are supposed to be redirected are instead sent out the default route. If this happens, a TCP ACK from the client host may be sent to the origin server, instead of being diverted to the cache. The origin server, upon receiving an unexpected ACK packet, sends a TCP RESET back to the client, which aborts the client's request.
To work around this problem, you can install a static route to the null0 interface for the cache address with a higher metric (lower precedence), such as 250. Then, when the link goes down, packets from the client just get dropped instead of sent out the default route. For example, if 1.2.3.4 is the IP address of your Squid cache, you may add:
ip route 1.2.3.4 255.255.255.255 Null0 250This appears to cause the correct behaviour.
Contributors: Glenn Chisholm, Lincoln Dale and Reuben Farrelly.
CISCO's Web Cache Coordination Protocol V1.0 is supported in squid 2.3 and later. support WCCP V2.0. Now that WCCP V2 is an open protocol, Squid may be able to support it in the future.
There are two different methods of configuring WCCP on CISCO routers. The first method is for routers that only support V1.0 of the protocol. The second is for routers that support both.
It is possible that later versions of IOS 11.x will support V2.0 of the protocol. If that is the case follow the 12.x instructions. Several people have reported that the squid implimentation of WCCP does not work with their 11.x routers. If you experience this please mail the debug output from your router to squid-bugs.
conf t wccp enable ! interface [Interface carrying Outgoing Traffic]x/x ! ip wccp web-cache redirect ! CTRL Z write mem
Some of the early versions of 12.x do not have the 'ip wccp version' command. You will need to upgrade your IOS version to use V1.0.
You will need to be running at least IOS Software Release 12.0(5)T if you're running the 12.0 T-train. IOS Software Releases 12.0(3)T and 12.0(4)T do not have WCCPv1, but 12.0(5)T does.
conf t ip wccp version 1 ip wccp web-cache redirect-list 150 ! interface [Interface carrying Outgoing/Incoming Traffic]x/x ip wccp web-cache redirect out|in ! CTRL Z write mem
Replace 150 with an access list number (either standard or extended) which lists IP addresses which you do not wish to be transparently redirected to your cache. Otherwise simply user the word 'redirect' on it's own to redirect traffic from all sources to all destinations.
Some people report problems with WCCP and IOS 12.x. They see truncated or fragmented GRE packets arriving at the cache. Apparently it works if you disable Cisco Express Forwarding for the interface:
conf t ip cef # some systems may already have 'ip cef global' int Ethernet 0/0 (or int FastEthernet 0/0 or other internal interface) no ip route-cache cef CTRL Z
This may well be fixed in later releases of IOS.
FreeBSD first needs to be configured to receive and strip the GRE encapsulation from the packets from the router. To do this you will need to patch and recompile your kernel. The steps depend on your kernel version.
# cd /usr/src # patch -s < /tmp/gre.patch
The procedure is nearly identical to the above for 3.x, but the source files are a little different.
The operating system now comes standard with some GRE support. You need to make a kernel with the GRE code enabled:
pseudo-device gre
And then configure the tunnel so that the router's GRE packets are accepted:
# ifconfig gre0 create # ifconfig gre0 $squid_ip $router_ip netmask 255.255.255.255 up # ifconfig gre0 tunnel $squid_ip $router_ip # route delete $router_ip
Alternatively, you can try it like this:
ifconfig gre0 create ifconfig gre0 $squid_ip 10.20.30.40 netmask 255.255.255.255 link1 tunnel $squid_ip $router_ip up
Since the WCCP/GRE tunnel is one-way, Squid never sends any packets to 10.20.30.40 and that particular address doesn't matter.
Al Blake has written a Cookbook for setting up transparent WCCP using Squid on RedHat Linux and a cisco access server.
There are currently two methods for supporting WCCP with Linux 2.2. A specific purpose module. Or the standard Linux GRE tunneling driver. People have reported difficulty with the standard GRE tunneling driver, however it does allow GRE functionality other than WCCP. You should choose the method that suits your enviroment.
Linux 2.2 kernels already support GRE, as long as the GRE module is compiled into the kernel. However, WCCP uses a slightly non-standard GRE encapsualtion format and Linux versions earlier than 2.6.9 may need to be patched to support WCCP.
Ensure that the GRE code is either built as static or as a module by chosing the appropriate option in your kernel config. Then rebuild your kernel. If it is a module you will need to:
modprobe ip_gre
The next step is to tell Linux to establish an IP tunnel between the router and your host. Daniele Orlandi reports that you have to give the gre1 interface an address, but any old address seems to work.
iptunnel add gre1 mode gre remote <Router-IP> local <Host-IP> dev <interface> ifconfig gre1 127.0.0.2 up<Router-IP> is the IP address of your router that is intercepting the HTTP packets. <Host-IP> is the IP address of your cache, and <interface> is the network interface that receives those packets (probably eth0).
Note that WCCP is incompatible with the rp_filter function in Linux and you must disable this if enabled (default off). If enabled any packets redirected by WCCP and intercepted by Netfilter/iptables will be silendly discarded by the TCP/IP stack due to their "unexpected" origin from the gre interface.
Joe Cooper has a patch for Linux 2.2.18 kernel on his Squid page.
This module is not part of the standard Linux distributon. It needs to be compiled as a module and loaded on your system to function. Do not attempt to build this in as a static part of your kernel.
Download the Linux WCCP module and compile it as you would any Linux network module.
Copy the module to /lib/modules/kernel-version/ipv4/ip_wccp.o. Edit /lib/modules/kernel-version/modules.dep and add:
/lib/modules/kernel-version/ipv4/ip_wccp.o:
or run
depmod -a
Finally you will need to load the module:
modprobe ip_wccp
The machine should now be striping the GRE encapsulation from any packets recieved and requeuing them. The system will also need to be configured for interception proxying, either with ipfwadm or with ipchains.
If you have managed to configuring your operating system to support WCCP with Squid please contact us with the details so we may share them with others.
IOS releases:
Cisco has published WCCPv2 as an Internet Draft (expired Jan 2001). There is a ongoing project at the Squid development projects website aiming to add support for WCCPv2 and at the time of writing this patch provides at least the same functionality as WCCPv1.
No, you cannot. With interception proxying, the client thinks it is talking to an origin server and would never send the Proxy-authorization request header.
by Joshua N Pritikin
#!/bin/sh iptables -t nat -F # clear table # normal transparent proxy iptables -t nat -A PREROUTING -p tcp -i eth0 --dport 80 -j REDIRECT --to-port 8080 # handle connections on the same box (192.168.0.2 is a loopback instance) gid=`id -g proxy` iptables -t nat -A OUTPUT -p tcp --dport 80 -m owner --gid-owner $gid -j ACCEPT iptables -t nat -A OUTPUT -p tcp --dport 80 -j DNAT --to-destination 192.168.0.2:8080
by Pedro A M Vazquez
On the switch define a network group to be intercepted:
policy network group MyGroup 10.1.1.0 mask 255.255.255.0
Define the tcp services to be intercepted:
policy service web80 destination tcp port 80 policy service web8080 destination tcp port 8080
Define a group of services using the services above:
policy service group WebPorts web80 web8080
And use these to create an intercept condition:
policy condition WebFlow source network group MyGroup service group WebPorts
Now, define an action to redirect the traffic to the host running squid:
policy action Redir alternate gateway ip 10.1.2.3
Finally, create a rule using this condition and the corresponding action:
policy rule Intercept condition WebFlow action Redir
Apply the rules to the QoS system to make them effective
qos apply
Don't forget that you still need to configure Squid and Squid's operating system to handle the intercepted connections. See above for Squid and OS-specific details.
Contributors: Glenn Chisholm.
True SNMP support is available in squid 2 and above. A significant change in the implimentation occured starting with the development 2.2 code. Therefore there are two sets of instructions on how to configure SNMP in squid, please make sure that you follow the correct one.
To use SNMP, it must first be enabled with the configure script, and squid rebuilt. To enable is first run the script:
./configure --enable-snmp [ ... other configure options ]Next, recompile after cleaning the source tree :
make clean make all make installOnce the compile is completed and the new binary is installed the squid.conf file needs to be configured to allow access; the default is to deny all requests. The instructions on how to do this have been broken into two parts, the first for all versions of Squid from 2.2 onwards and the second for 2.1 and below.
To configure SNMP first specify a list of communities that you would like to allow access by using a standard acl of the form:
acl aclname snmp_community stringFor example:
acl snmppublic snmp_community public acl snmpjoebloggs snmp_community joebloggsThis creates two acl's, with two different communities, public and joebloggs. You can name the acl's and the community strings anything that you like.
To specify the port that the agent will listen on modify the "snmp_port" parameter, it is defaulted to 3401. The port that the agent will forward requests that can not be furfilled by this agent to is set by "forward_snmpd_port" it is defaulted to off. It must be configured for this to work. Remember that as the requests will be originating from this agent you will need to make sure that you configure your access accordingly.
To allow access to Squid's SNMP agent, define an snmp_access ACL with the community strings that you previously defined. For example:
snmp_access allow snmppublic localhost snmp_access deny allThe above will allow anyone on the localhost who uses the community public to access the agent. It will deny all others access.
If you do not define any snmp_access ACL's, then SNMP access is denied by default.
Finally squid allows to you to configure the address that the agent will bind to for incomming and outgoing traffic. These are defaulted to 0.0.0.0, changing these will cause the agent to bind to a specific address on the host, rather than the default which is all.
snmp_incoming_address 0.0.0.0 snmp_outgoing_address 0.0.0.0
Prior to Squid 2.1 the SNMP code had a number of issues with the ACL's. If you are a frequent user of SNMP with Squid, please upgrade to 2.2 or higher.
A sort of default, working configuration is:
snmp_port 3401 snmp_mib_path /local/squid/etc/mib.txt snmp_agent_conf view all .1.3.6 included snmp_agent_conf view squid .1.3.6 included snmp_agent_conf user squid - all all public snmp_agent_conf user all all all all squid snmp_agent_conf community public squid squid snmp_agent_conf community readwrite all all
Note that for security you are advised to restrict SNMP access to your caches. You can do this easily as follows:
acl snmpmanagementhosts 1.2.3.4/255.255.255.255 1.2.3.0/255.255.255.0 snmp_acl public deny all !snmpmanagementhosts snmp_acl readwrite deny allYou must follow these instructions for 2.1 and below exactly or you are likely to have problems. The parser has some issues which have been corrected in 2.2.
You can test if your Squid supports SNMP with the snmpwalk program (snmpwalk is a part of the NET-SNMP project). Note that you have to specify the SNMP port, which in Squid defaults to 3401.
snmpwalk -p 3401 hostname communitystring .1.3.6.1.4.1.3495.1.1If it gives output like:
enterprises.nlanr.squid.cacheSystem.cacheSysVMsize = 7970816 enterprises.nlanr.squid.cacheSystem.cacheSysStorage = 2796142 enterprises.nlanr.squid.cacheSystem.cacheUptime = Timeticks: (766299) 2:07:42.99then it is working ok, and you should be able to make nice statistics out of it.
For an explanation of what every string (OID) does, you should refer to the Squid SNMP web pages.
There are a lot of things you can do with SNMP and Squid. It can be useful in some extent for a longer term overview of how your proxy is doing. It can also be used as a problem solver. For example: how is it going with your filedescriptor usage? or how much does your LRU vary along a day. Things you can't monitor very well normally, aside from clicking at the cachemgr frequently. Why not let MRTG do it for you?
There are a number of tools that you can use to monitor Squid via SNMP. Many people use MRTG. Another good combination is NET-SNMP plus RRDTool. You might be able to find more information at the Squid SNMP web pages or ircache rrdtool scipts
General Discussion: cache-snmp@ircache.net These messages are archived.
Subscriptions should be sent to: cache-snmp-request@ircache.net.
Some people use MRTG to query Squid through its SNMP interface.
To get instruction on using MRTG with Squid please visit these pages:
Further examples of Squid MRTG configurations can be found here:
By default, Squid connects directly to origin servers for SSL requests. But if you must force SSL requests through a parent, first tell Squid it can not go direct for SSL:
acl SSL method CONNECT never_direct allow SSLWith this in place, Squid should pick one of your parents to use for SSL requests. If you want it to pick a particular parent, you must use the cache_peer_access configuration:
cache_peer parent1 parent 3128 3130 cache_peer parent2 parent 3128 3130 cache_peer_access parent2 allow !SSLThe above lines tell Squid to NOT use parent2 for SSL, so it should always use parent1.
Simply add your new cache_dir line to squid.conf, then run squid -z again. Squid will create swap directories on the new disk and leave the existing ones in place.
Authentication is handled via external processes. Arjan's proxy auth page describes how to set it up. Some simple instructions are given below as well.
acl foo proxy_auth REQUIRED http_access allow foo
% cd helpers/basic_auth/NCSA % make % make installYou should now have an ncsa_auth program in the <prefix>/libexec/ directory where the helpers for squid lives (usually /usr/local/squid/libexec unless overridden by configure flags). You can also select with the --enable-basic-auth-helpers=... option which helpers should be installed by default when you install Squid.
auth_param basic program /usr/local/squid/libexec/ncsa_auth /usr/local/squid/etc/passwd
After all that, you should be able to start up Squid. If we left something out, or haven't been clear enough, please let us know (squid-faq@squid-cache.org).
The ACL for proxy-authentication has changed from:
acl foo proxy_auth timeoutto:
acl foo proxy_auth usernamePlease update your ACL appropriately - a username of REQUIRED will permit all valid usernames. The timeout is now specified with the configuration option:
auth_param basic credentialsttl timeout
by David Luyer.
The information here is current for version 2.2. It is strongly recommended that you use at least Squid 2.2 if you wish to use delay pools.
Delay pools provide a way to limit the bandwidth of certain requests based on any list of criteria. The idea came from a Western Australian university who wanted to restrict student traffic costs (without affecting staff traffic, and still getting cache and local peering hits at full speed). There was some early Squid 1.0 code by Central Network Services at Murdoch University, which I then developed (at the University of Western Australia) into a much more complex patch for Squid 1.0 called ``DELAY_HACK.'' I then tried to code it in a much cleaner style and with slightly more generic options than I personally needed, and called this ``delay pools'' in Squid 2. I almost completely recoded this in Squid 2.2 to provide the greater flexibility requested by people using the feature.
To enable delay pools features in Squid 2.2, you must use the --enable-delay-pools configure option before compilation.
Terminology for this FAQ entry:
a collection of bucket groups as appropriate to a given class
a group of buckets within a pool, such as the per-host bucket group, the per-network bucket group or the aggregate bucket group (the aggregate bucket group is actually a single bucket)
an individual delay bucket represents a traffic allocation which is replenished at a given rate (up to a given limit) and causes traffic to be delayed when empty
the class of a delay pool determines how the delay is applied, ie, whether the different client IPs are treated seperately or as a group (or both)
a class 1 delay pool contains a single unified bucket which is used for all requests from hosts subject to the pool
a class 2 delay pool contains one unified bucket and 255 buckets, one for each host on an 8-bit network (IPv4 class C)
contains 255 buckets for the subnets in a 16-bit network, and individual buckets for every host on these networks (IPv4 class B)
Delay pools allows you to limit traffic for clients or client groups, with various features:
This allows options such as creating a number of class 1 delay pools and allowing a certain amount of bandwidth to given object types (by using URL regular expressions or similar), and many other uses I'm sure I haven't even though of beyond the original fair balancing of a relatively small traffic allocation across a large number of users.
There are some limitations of delay pools:
acl all src 0.0.0.0/0.0.0.0 # might already be defined delay_pools 1 delay_class 1 1 delay_access 1 allow all delay_parameters 1 64000/64000 # 512 kbits == 64 kbytes per second
For an explanation of these tags please see the configuration file.
The 1 second buffer (max = restore = 64kbytes/sec) is because a limit is requested, and no responsiveness to a busrt is requested. If you want it to be able to respond to a burst, increase the aggregate_max to a larger value, and traffic bursts will be handled. It is recommended that the maximum is at least twice the restore value - if there is only a single object being downloaded, sometimes the download rate will fall below the requested throughput as the bucket is not empty when it comes to be replenished.
You can not limit a single HTTP request's connection speed. You can limit individual hosts to some bandwidth rate. To limit a specific host, define an acl for that host and use the example above. To limit a group of hosts, then you must use a delay pool of class 2 or 3. For example:
acl only128kusers src 192.168.1.0/255.255.192.0 acl all src 0.0.0.0/0.0.0.0 delay_pools 1 delay_class 1 3 delay_access 1 allow only128kusers delay_access 1 deny all delay_parameters 1 64000/64000 -1/-1 16000/64000For an explanation of these tags please see the configuration file.
The above gives a solution where a cache is given a total of 512kbits to operate in, and each IP address gets only 128kbits out of that pool.
We have six local cache peers, all with the options 'proxy-only no-delay' since they are fast machines connected via a fast ethernet and microwave (ATM) network.
For our local access we use a dstdomain ACL, and for delay pool exceptions we use a dst ACL as well since the delay pool ACL processing is done using "fast lookups", which means (among other things) it won't wait for a DNS lookup if it would need one.
Our proxy has two virtual interfaces, one which requires student authentication to connect from machines where a department is not paying for traffic, and one which uses delay pools. Also, users of the main Unix system are allowed to choose slow or fast traffic, but must pay for any traffic they do using the fast cache. Ident lookups are disabled for accesses through the slow cache since they aren't needed. Slow accesses are delayed using a class 3 delay pool to give fairness between departments as well as between users. We recognize users of Lynx on the main host are grouped together in one delay bucket but they are mostly viewing text pages anyway, so this isn't considered a serious problem. If it was we could take those hosts into a class 1 delay pool and give it a larger allocation.
I prefer using a slow restore rate and a large maximum rate to give preference to people who are looking at web pages as their individual bucket fills while they are reading, and those downloading large objects are disadvantaged. This depends on which clients you believe are more important. Also, one individual 8 bit network (a residential college) have paid extra to get more bandwidth.
The relevant parts of my configuration file are (IP addresses, etc, all changed):
# ACL definitions # Local network definitions, domains a.net, b.net acl LOCAL-NET dstdomain a.net b.net # Local network; nets 64 - 127. Also nearby network class A, 10. acl LOCAL-IP dst 192.168.64.0/255.255.192.0 10.0.0.0/255.0.0.0 # Virtual i/f used for slow access acl virtual_slowcache myip 192.168.100.13/255.255.255.255 # All permitted slow access, nets 96 - 127 acl slownets src 192.168.96.0/255.255.224.0 # Special 'fast' slow access, net 123 acl fast_slow src 192.168.123.0/255.255.255.0 # User hosts acl my_user_hosts src 192.168.100.2/255.255.255.254 # "All" ACL acl all src 0.0.0.0/0.0.0.0 # Don't need ident lookups for billing on (free) slow cache ident_lookup_access allow my_user_hosts !virtual_slowcache ident_lookup_access deny all # Security access checks http_access [...] # These people get in for slow cache access http_access allow virtual_slowcache slownets http_access deny virtual_slowcache # Access checks for main cache http_access [...] # Delay definitions (read config file for clarification) delay_pools 2 delay_initial_bucket_level 50 delay_class 1 3 delay_access 1 allow virtual_slowcache !LOCAL-NET !LOCAL-IP !fast_slow delay_access 1 deny all delay_parameters 1 8192/131072 1024/65536 256/32768 delay_class 2 2 delay_access 2 allow virtual_slowcache !LOCAL-NET !LOCAL-IP fast_slow delay_access 2 deny all delay_parameters 2 2048/65536 512/32768
The same code is also used by a some of departments using class 2 delay pools to give them more flexibility in giving different performance to different labs or students.
This is also pretty well documented in the configuration file, with examples. Since people seem to lose their config files, here's a copy of the relevant section.
# DELAY POOL PARAMETERS (all require DELAY_POOLS compilation option) # ----------------------------------------------------------------------------- # TAG: delay_pools # This represents the number of delay pools to be used. For example, # if you have one class 2 delay pool and one class 3 delays pool, you # have a total of 2 delay pools. # # To enable this option, you must use --enable-delay-pools with the # configure script. #delay_pools 0 # TAG: delay_class # This defines the class of each delay pool. There must be exactly one # delay_class line for each delay pool. For example, to define two # delay pools, one of class 2 and one of class 3, the settings above # and here would be: # #delay_pools 2 # 2 delay pools #delay_class 1 2 # pool 1 is a class 2 pool #delay_class 2 3 # pool 2 is a class 3 pool # # The delay pool classes are: # # class 1 Everything is limited by a single aggregate # bucket. # # class 2 Everything is limited by a single aggregate # bucket as well as an "individual" bucket chosen # from bits 25 through 32 of the IP address. # # class 3 Everything is limited by a single aggregate # bucket as well as a "network" bucket chosen # from bits 17 through 24 of the IP address and a # "individual" bucket chosen from bits 17 through # 32 of the IP address. # # NOTE: If an IP address is a.b.c.d # -> bits 25 through 32 are "d" # -> bits 17 through 24 are "c" # -> bits 17 through 32 are "c * 256 + d" # TAG: delay_access # This is used to determine which delay pool a request falls into. # The first matched delay pool is always used, ie, if a request falls # into delay pool number one, no more delay are checked, otherwise the # rest are checked in order of their delay pool number until they have # all been checked. For example, if you want some_big_clients in delay # pool 1 and lotsa_little_clients in delay pool 2: # #delay_access 1 allow some_big_clients #delay_access 1 deny all #delay_access 2 allow lotsa_little_clients #delay_access 2 deny all # TAG: delay_parameters # This defines the parameters for a delay pool. Each delay pool has # a number of "buckets" associated with it, as explained in the # description of delay_class. For a class 1 delay pool, the syntax is: # #delay_parameters pool aggregate # # For a class 2 delay pool: # #delay_parameters pool aggregate individual # # For a class 3 delay pool: # #delay_parameters pool aggregate network individual # # The variables here are: # # pool a pool number - ie, a number between 1 and the # number specified in delay_pools as used in # delay_class lines. # # aggregate the "delay parameters" for the aggregate bucket # (class 1, 2, 3). # # individual the "delay parameters" for the individual # buckets (class 2, 3). # # network the "delay parameters" for the network buckets # (class 3). # # A pair of delay parameters is written restore/maximum, where restore is # the number of bytes (not bits - modem and network speeds are usually # quoted in bits) per second placed into the bucket, and maximum is the # maximum number of bytes which can be in the bucket at any time. # # For example, if delay pool number 1 is a class 2 delay pool as in the # above example, and is being used to strictly limit each host to 64kbps # (plus overheads), with no overall limit, the line is: # #delay_parameters 1 -1/-1 8000/8000 # # Note that the figure -1 is used to represent "unlimited". # # And, if delay pool number 2 is a class 3 delay pool as in the above # example, and you want to limit it to a total of 256kbps (strict limit) # with each 8-bit network permitted 64kbps (strict limit) and each # individual host permitted 4800bps with a bucket maximum size of 64kb # to permit a decent web page to be downloaded at a decent speed # (if the network is not being limited due to overuse) but slow down # large downloads more significantly: # #delay_parameters 2 32000/32000 8000/8000 600/8000 # # There must be one delay_parameters line for each delay pool. # TAG: delay_initial_bucket_level (percent, 0-100) # The initial bucket percentage is used to determine how much is put # in each bucket when squid starts, is reconfigured, or first notices # a host accessing it (in class 2 and class 3, individual hosts and # networks only have buckets associated with them once they have been # "seen" by squid). # #delay_initial_bucket_level 50
At the moment we do not have a script which will convert your cache contents from the 1.1 to the Squid-2 format. If enough people ask for one, then somebody will probably write such a script.
If you like, you can configure a new Squid-2 cache with your old Squid-1.1 cache as a sibling. After a few days, weeks, or however long you want to wait, shut down the old Squid cache. If you want to force-load your new cache with the objects from the old cache, you can try something like this:
Squid-2 lets you customize your error messages. The source distribution includes error messages in different languages. You can select the language with the configure option:
--enable-err-language=lang
Furthermore, you can rewrite the error message template files if you like. This list describes the tags which Squid will insert into the messages:
URL with FTP %2f hack
Squid error code
seconds elapsed since request received (not yet implemented)
errno
strerror()
FTP request line
FTP reply line
FTP server message
cache hostname
server host name
client IP address
server IP address
contents of err_html_text config option
Request Method
Error message returned by external auth helper
URL port \#
Protocol
Full HTTP Request
squid default signature
caching proxy software with version
local time
UTC
URL without password
URL with password (Squid-2.5 and later only)
cachemgr email address
dns server error message
The Squid default signature is added automatically unless %s is used in the error page. To change the signature you must manually append the signature to each error page.
The default signature reads like:
<BR clear="all">
<HR noshade size="1px">
<ADDRESS>
Generated %T by %h (%s)
</ADDRESS>
</BODY></HTML>
Yes, a number of configuration directives have been renamed. Here are some of them:
This is now called cache_peer. The old term does not really describe what you are configuring, but the new name tells you that you are configuring a peer for your cache.
Renamed to cache_peer_domain.
The functaionality provided by these directives is now implemented as access control lists. You will use the always_direct and never_direct options. The new squid.conf file has some examples.
This directive also has been reimplemented with access control lists. You will use the no_cache option. For example:
acl Uncachable url_regex cgi ? no_cache deny Uncachable
This option used to specify the cache disk size. Now you specify the disk size on each cache_dir line.
This option has been renamed to cache_peer_access and the syntax has changed. Now this option is a true access control list, and you must include an allow or deny keyword. For example:
acl that-AS dst_as 1241 cache_peer_access thatcache.thatdomain.net allow that-AS cache_peer_access thatcache.thatdomain.net deny allThis example sends requests to your peer thatcache.thatdomain.net only for origin servers in Autonomous System Number 1241.
In Squid-1.1 many of the configuration options had implied units associated with them. For example, the connect_timeout value may have been in seconds, but the read_timeout value had to be given in minutes. With Squid-2, these directives take units after the numbers, and you will get a warning if you leave off the units. For example, you should now write:
connect_timeout 120 seconds read_timeout 15 minutes
Occasionally people have trouble understanding accelerators and proxy caches, usually resulting from mixed up interpretations of "incoming" and ``outgoing" data. I think in terms of requests (i.e., an outgoing request is from the local site out to the big bad Internet). The data received in reply is incoming, of course. Others think in the opposite sense of ``a request for incoming data".
An accelerator caches incoming requests for outgoing data (i.e., that which you publish to the world). It takes load away from your HTTP server and internal network. You move the server away from port 80 (or whatever your published port is), and substitute the accelerator, which then pulls the HTTP data from the ``real" HTTP server (only the accelerator needs to know where the real server is). The outside world sees no difference (apart from an increase in speed, with luck).
Quite apart from taking the load of a site's normal web server, accelerators can also sit outside firewalls or other network bottlenecks and talk to HTTP servers inside, reducing traffic across the bottleneck and simplifying the configuration. Two or more accelerators communicating via ICP can increase the speed and resilience of a web service to any single failure.
The Squid redirector can make one accelerator act as a single front-end for multiple servers. If you need to move parts of your filesystem from one server to another, or if separately administered HTTP servers should logically appear under a single URL hierarchy, the accelerator makes the right thing happen.
If you wish only to cache the ``rest of the world" to improve local users browsing performance, then accelerator mode is irrelevant. Sites which own and publish a URL hierarchy use an accelerator to improve other sites' access to it. Sites wishing to improve their local users' access to other sites' URLs use proxy caches. Many sites, like us, do both and hence run both.
Measurement of the Squid cache and its Harvest counterpart suggest an order of magnitude performance improvement over CERN or other widely available caching software. This order of magnitude performance improvement on hits suggests that the cache can serve as an httpd accelerator, a cache configured to act as a site's primary httpd server (on port 80), forwarding references that miss to the site's real httpd (on port 81).
In such a configuration, the web administrator renames all non-cachable URLs to the httpd's port (81). The cache serves references to cachable objects, such as HTML pages and GIFs, and the true httpd (on port 81) serves references to non-cachable objects, such as queries and cgi-bin programs. If a site's usage characteristics tend toward cachable objects, this configuration can dramatically reduce the site's web workload.
Note that it is best not to run a single squid process as
both an httpd-accelerator and a proxy cache, since these two modes
will have different working sets. You will get better performance
by running two separate caches on separate machines. However, for
compatability with how administrators are accustomed to running
other servers that provide both proxy and Web serving capability
(eg, CERN), the Squid supports operation as both a proxy and
an accelerator if you set the httpd_accel_with_proxy
variable to on inside your squid.conf
configuration file.
First, you have to tell Squid to listen on port 80 (usually), so set the 'http_port' option:
http_port 80
Next, you need to move your normal HTTP server to another port and/or another machine. If you want to run your HTTP server on the same machine, then it can not also use port 80 (except see the next FAQ entry below). A common choice is port 81. Configure squid as follows:
httpd_accel_host localhost httpd_accel_port 81Alternatively, you could move the HTTP server to another machine and leave it on port 80:
httpd_accel_host otherhost.foo.com httpd_accel_port 80
You should now be able to start Squid and it will serve requests as a HTTP server.
If you are using Squid has an accelerator for a virtual host system, then you need to specify
httpd_accel_host virtual
Finally, if you want Squid to also accept proxy requests (like it used to before you turned it into an accelerator), then you need to enable this option:
httpd_accel_with_proxy on
Yes, this is because you probably moved your real httpd to port 81. When your httpd issues a redirect message (e.g. 302 Moved Temporarily), it knows it is not running on the standard port (80), so it inserts :81 in the redirected URL. Then, when the client requests the redirected URL, it bypasses the accelerator.
How can you fix this?
One way is to leave your httpd running on port 80, but bind the httpd socket to a specific interface, namely the loopback interface. With Apache you can do it like this in httpd.conf:
Port 80 BindAddress 127.0.0.1Then, in your squid.conf file, you must specify the loopback address as the accelerator:
httpd_accel_host 127.0.0.1 httpd_accel_port 80
Note, you probably also need to add an /etc/hosts entry of 127.0.0.1 for your server hostname. Otherwise, Squid may get stuck in a forwarding loop.
Wget is a command-line Web client. It supports HTTP and FTP URLs, recursive retrievals, and HTTP proxies.
If you want to test your Squid cache in batch (from a cron command, for instance), you can use the echoping program, which will tell you (in plain text or via an exit code) if the cache is up or not, and will indicate the response times.
Rather than maintain the same list in two places, please see the Logfile Analysis Scripts page on the Web server.
Kenichi Matsui has a simple perl script which generates a 3D hierarchy map (in VRML) from squid.conf. 3Dhierarchy.pl.
transproxy is a program used in conjunction with the Linux Transparent Proxy networking feature, and ipfwadm, to intercept HTTP and other requests. Transproxy is written by John Saunders.
A redirector package from Iain Lea to allow Intranet (restricted) or Internet (full) access with URL deny and redirection for sites that are not deemed acceptable for a userbase all via a single proxy port.
Junkbusters Corp has a copyleft privacy-enhancing, ad-blocking proxy server which you can use in conjunction with Squid.
Squirm is a configurable, efficient redirector for Squid by Chris Foote. Features:
Pedro L Orso has adapated the Apache's htpasswd into a CGI program called chpasswd.cgi.
jesred by Jens Elkner.
squidGuard is a free (GPL), flexible and efficient filter and redirector program for squid. It lets you define multiple access rules with different restrictions for different user groups on a squid cache. squidGuard uses squid standard redirector interface.
The Smart Neighbour [URL disappeared] (or 'Central Squid Server' - CSS) is a cut-down version of Squid without HTTP or object caching functionality. The CSS deals only with ICP messages. Instead of caching objects, the CSS records the availability of objects in each of its neighbour caches. Caches that have smart neighbours update each smart neighbour with the status of their cache by sending ICP_STORE_NOTIFY/ICP_RELEASE_NOTIFY messages upon storing/releasing an object from their cache. The CSS maintains an up to date 'object map' recording the availability of objects in its neighbouring caches.
The Cerberian content filter is a very flexible URL rating system with full Squid integration provided by MARA Systems AB. The service requires a license (priced by the number of seats) but evaluation licenses are available.
For Windows NT, Windows 95/98, and Unix.
DISKD refers to some features in Squid-2.4 and later to improve Disk I/O performance. The basic idea is that each cache_dir has its own diskd child process. The diskd process performs all disk I/O operations (open, close, read, write, unlink) for the cache_dir. Message queues are used to send requests and responses between the Squid and diskd processes. Shared memory is used for chunks of data to be read and written.
Yes. We benchmarked Squid-2.4 with DISKD at the Second IRCache Bake-Off. The results are also described here. At the bakeoff, we got 160 req/sec with diskd. Without diskd, we'd have gotten about 40 req/sec.
You need to run Squid version 2.4 or later. Your operating system must support message queues, and shared memory.
To configure Squid for DISKD, use the --enable-storeio option:
% ./configure --enable-storeio=diskd,ufs
You didn't put diskd in the list of storeio modules as described above. You need to run configure and and recompile Squid.
No. Diskd uses the same storage scheme as the standard "UFS" type. It only changes how I/O is performed.
Most Unix operating systems have message queue support by default. One way to check is to see if you have an ipcs command.
However, you will likely need to increase the message queue parameters for Squid. Message queue implementations normally have the following parameters:
Maximum number of bytes per message queue.
Maximum number of message queue identifiers (system wide).
Maximum number of message segments per queue.
Size of a message segment.
Maximum number of messages (system wide).
Maximum size of a whole message. On some systems you may need to increase this limit. On other systems, you may not be able to change it.
The messages between Squid and diskd are 32 bytes for 32-bit CPUs and 40 bytes for 64-bit CPUs. Thus, MSGSSZ should be 32 or greater. You may want to set it to a larger value, just to be safe.
We'll have two queues for each cache_dir -- one in each direction. So, MSGMNI needs to be at least two times the number of cache_dir's.
I've found that 75 messages per queue is about the limit of decent performance. If each diskd message consists of just one segment (depending on your value of MSGSSZ), then MSGSEG should be greater than 75.
MSGMNB and MSGTQL affect how many messages can be in the queues at one time. Diskd messages shouldn't be more than 40 bytes, but let's use 64 bytes to be safe. MSGMNB should be at least 64*75. I recommend rounding up to the nearest power of two, or 8192.
MSGTQL should be at least 75 times the number of cache_dir's that you'll have.
Your kernel must have
options SYSVMSG
You can set the parameters in the kernel as follows. This is just an example. Make sure the values are appropriate for your system:
options MSGMNB=8192 # max # of bytes in a queue options MSGMNI=40 # number of message queue identifiers options MSGSEG=512 # number of message segments per queue options MSGSSZ=64 # size of a message segment options MSGTQL=2048 # max messages in system
You can set the parameters in the kernel as follows. This is just an example. Make sure the values are appropriate for your system:
option MSGMNB=16384 # max characters per message queue option MSGMNI=40 # max number of message queue identifiers option MSGSEG=2048 # max number of message segments in the system option MSGSSZ=64 # size of a message segment (Must be 2^N) option MSGTQL=1024 # max amount of messages in the system
Message queue support seems to be in the kernel by default. Setting the options is as follows:
options MSGMNB="8192" # max # bytes on queue options MSGMNI="40" # # of message queue identifiers options MSGMAX="2048" # max message size options MSGTQL="2048" # # of system message headers
If you have a newer version (DU64), then you can probably use sysconfig instead. To see what the current IPC settings are run
# sysconfig -q ipcTo change them make a file like this called ipc.stanza:
ipc:
msg-max = 2048
msg-mni = 40
msg-tql = 2048
msg-mnb = 8192
then run
# sysconfigdb -a -f ipc.stanzaYou have to reboot for the change to take effect.
Stefan Köpsell reports that if you compile sysctl support into your kernel, then you can change the following values:
Winfried Truemper reports: The default values should be large enough for most common cases. You can modify the message queue configuration by writing to these files:
Refer to Demangling Message Queues in Sunworld Magazine.
I don't think the above article really tells you how to set the parameters. You do it in /etc/system with lines like this:
set msgsys:msginfo_msgmax=2048 set msgsys:msginfo_msgmnb=8192 set msgsys:msginfo_msgmni=40 set msgsys:msginfo_msgssz=64 set msgsys:msginfo_msgtql=2048
Of course, you must reboot whenever you modify /etc/system before changes take effect.
Shared memory uses a set of parameters similar to the ones for message queues. The Squid DISKD implementation uses one shared memory area for each cache_dir. Each shared memory area is about 800 kilobytes in size. You may need to modify your system's shared memory parameters:
Maximum number of shared memory segments per process.
Maximum number of shared memory segments for the whole system.
Largest shared memory segment size allowed.
Total amount of shared memory that can be used.
For Squid and DISKD, SHMMNI and SHMMNI must be greater than or equal to the number of cache_dir's that you have. SHMMAX must be at least 800 kilobytes. SHMALL must be at least SHMMAX 800 kilobytes multiplied by the number of cache_dir's.
Your kernel must have
options SYSVSHM
You can set the parameters in the kernel as follows. This is just an example. Make sure the values are appropriate for your system:
options SHMSEG=16 # max shared mem id's per process options SHMMNI=32 # max shared mem id's per system options SHMMAX=2097152 # max shared memory segment size (bytes) options SHMALL=4096 # max amount of shared memory (pages)
OpenBSD is similar to FreeBSD, except you must use option instead of options, and SHMMAX is in pages instead of bytes:
option SHMSEG=16 # max shared mem id's per process option SHMMNI=32 # max shared mem id's per system option SHMMAX=2048 # max shared memory segment size (pages) option SHMALL=4096 # max amount of shared memory (pages)
Message queue support seems to be in the kernel by default. Setting the options is as follows:
options SHMSEG="16" # max shared mem id's per process options SHMMNI="32" # max shared mem id's per system options SHMMAX="2097152" # max shared memory segment size (bytes) options SHMALL=4096 # max amount of shared memory (pages)
If you have a newer version (DU64), then you can probably use sysconfig instead. To see what the current IPC settings are run
# sysconfig -q ipcTo change them make a file like this called ipc.stanza:
ipc:
shm-seg = 16
shm-mni = 32
shm-max = 2097152
shm-all = 4096
then run
# sysconfigdb -a -f ipc.stanzaYou have to reboot for the change to take effect.
Winfried Truemper reports: The default values should be large enough for most common cases. You can modify the shared memory configuration by writing to these files:
Stefan Köpsell reports that if you compile sysctl support into your kernel, then you can change the following values:
Refer to Shared memory uncovered in Sunworld Magazine.
To set the values, you can put these lines in /etc/system:
set shmsys:shminfo_shmmax=2097152 set shmsys:shminfo_shmmni=32 set shmsys:shminfo_shmseg=16
Yes, this is a little problem sometimes. Seems like the operating system gets confused and doesn't always release shared memory and message queue resources when processes exit, especially if they exit abnormally. To fix it you can ``manually'' clear the resources with the ipcs command. Add this command into your RunCache or squid_start script:
ipcs | grep '^[mq]' | awk '{printf "ipcrm -%s %s\n", $1, $2}' | /bin/sh
In the source code, these are called magic1 and magic2. These numbers refer to the number of oustanding requests on a message queue. They are specified on the cache_dir option line, after the L1 and L2 directories:
cache_dir diskd /cache1 1024 16 256 Q1=72 Q2=64
If there are more than Q1 messages outstanding, then Squid will intentionally fail to open disk files for reading and writing. This is a load-shedding mechanism. If your cache gets really really busy and the disks can not keep up, Squid bypasses the disks until the load goes down again.
If there are more than Q2 messages outstanding, then the main Squid process ``blocks'' for a little bit until the diskd process services some of the messages and sends back some replies.
Reasonable Q1 and Q2 values are 64 and 72. If you would rather have good hit ratio and bad response time, set Q1 > Q2. Otherwise, if you would rather have good response time and bad hit ratio, set Q1 < Q2.
Note: The information here is current for version 2.5.
Users will be authenticated if squid is configured to use proxy_auth ACLs (see next question).
Browsers send the user's authentication credentials in the Authorization request header.
If Squid gets a request and the http_access rule list gets to a proxy_auth ACL, Squid looks for the Authorization header. If the header is present, Squid decodes it and extracts a username and password.
If the header is missing, Squid returns an HTTP reply with status 407 (Proxy Authentication Required). The user agent (browser) receives the 407 reply and then prompts the user to enter a name and password. The name and password are encoded, and sent in the Authorization header for subsequent requests to the proxy.
NOTE: The name and password are encoded using ``base64'' (See section 11.1 of RFC 2616). However, base64 is a binary-to-text encoding only, it does NOT encrypt the information it encodes. This means that the username and password are essentially ``cleartext'' between the browser and the proxy. Therefore, you probably should not use the same username and password that you would use for your account login.
Authentication is actually performed outside of main Squid process. When Squid starts, it spawns a number of authentication subprocesses. These processes read usernames and passwords on stdin, and reply with "OK" or "ERR" on stdout. This technique allows you to use a number of different authentication schemes, although currently you can only use one scheme at a time.
The Squid source code comes with a few authentcation processes for Basic authentication. These include:
In addition Squid also supports the NTLM and Digest authentication schemes which both provide more secure authentication methods where the password is not exchanged in plain text. Each scheme have their own set of helpers and auth_param settings. You can not mix helpers between the different authentication schemes. For information on how to set up NTLM authentication see winbind .
In order to authenticate users, you need to compile and install one of the supplied authentication modules found in the helpers/basic_auth/ directory, one of the others, or supply your own.
You tell Squid which authentication program to use with the auth_param option in squid.conf. You specify the name of the program, plus any command line options if necessary. For example:
auth_param basic program /usr/local/squid/bin/ncsa_auth /usr/local/squid/etc/passwd
Make sure that your authentication program is installed and working correctly. You can test it by hand.
Add some proxy_auth ACL entries to your squid configuration. For example:
acl foo proxy_auth REQUIRED acl all src 0/0 http_access allow foo http_access deny allThe REQURIED term means that any authenticated user will match the ACL named foo.
Squid allows you to provide fine-grained controls by specifying individual user names. For example:
acl foo proxy_auth REQUIRED acl bar proxy_auth lisa sarah frank joe acl daytime time 08:00-17:00 acl all src 0/0 http_access allow bar http_access allow foo daytime http_access deny allIn this example, users named lisa, sarah, joe, and frank are allowed to use the proxy at all times. Other users are allowed only during daytime hours.
Yes. Successful authentication lookups are cached for one hour by default. That means (in the worst case) its possible for someone to keep using your cache up to an hour after he has been removed from the authentication database.
You can control the expiration time with the auth_param option.
Note: This has nothing to do with how often the user needs to re-authenticate himself. It is the browser who maintains the session, and re-authentication is a business between the user and his browser, not the browser and Squid. The browser authenticates on behalf of the user on every request sent to Squid. What this parameter controls is only how often Squid will ask the defined helper if the password is still valid.
Squid stores cleartext passwords in itsmemory cache.
Squid writes cleartext usernames and passwords when talking to the external authentication processes. Note, however, that this interprocess communication occors over TCP connections bound to the loopback interface or private UNIX pipes. Thus, its not possile for processes on other comuters or local users without root privileges to "snoop" on the authentication traffic.
Each authentication program must select its own scheme for persistent storage of passwords and usernames.
Winbind is a recent addition to Samba providing some impressive capabilities for NT based user accounts. From Squid's perspective winbind provides a robust and efficient engine for both basic and NTLM challenge/response authentication against an NT domain controller.
The winbind authenticators have been used successfully under Linux, FreeBSD and Solaris.
Samba-3.X is supported natively using the ntlm_auth helper shipped as part of Samba. No Squid specific winbind helpers need to be compiled (and even if compiled they won't work with Samba-3.X)
Samba-2.2.X is supported using the winbind helpers shipped with Squid, and uses an internal Samba interface to communicate with the winbindd daemon. It is therefore sensitive to any changes the Samba team may make to the interface.
The winbind helpers shipped with Squid-2.5.STABLE2 supports Samba-2.2.6 to Samba-2.2.7a and hopefully later Samba-2.X versions. To use Squid-2.5.STABLE2 with Samba versions 2.2.5 or ealier the new --with-samba-sources=... configure option is required. This may also be the case with Samba-2.2.X versions later than 2.2.7a or if you have applied any winbind related patches to your Samba tree.
Squid-2.5.STABLE1 supported Samba 2.2.4 or 2.2.5 only. Use of Squid-2.5.STABLE2 or later recommended with current Samba-2.X releases.
For Samba-3.X the winbind helpers shipped with Squid should not be used (and won't work if your attempt to do so), instead the ntlm_auth helper shipped as part of the Samba-3 distribution should be used. This helper supports all versions of Squid and both the ntlm and basic authentication schemes. For details on how to use this Samba helper see the Samba documentation. For group membership lookups the wbinfo_group helper shipped with Squid can be used (this is just a wrapper around the samba wbinfo program and works with all versions of Samba)
For full details on how to configure Samba and joining a domain please see the Samba documentation. The Samba team has quite extensive documentation both on how to join a NT domain and how to join a Active Directory tree.
Build/Install Samba-3.X
Samba must be built with these configure options:
--with-winbind
Then follow the Samba installation instructions. But please note that neither nsswitch or the pam modules needs to be installed for Squid to function, these are only needed if you want your OS to integrate with the domain for UNIX accounts.
Build/Install Samba-2.2.X
Samba must be built with these configure options:
--with-winbind --with-winbind-auth-challenge
Optionally, if building Samba 2.2.5, apply the smbpasswd.diff patch. See SMBD and Machine Trust Accounts below to determine if the patch is worthwhile.
Test Samba's winbindd
workgroup = mydomain password server = myPDC security = domain winbind uid = 10000-20000 winbind gid = 10000-20000 winbind use default domain = yes
# wbinfo -t Secret is good
# wbinfo -a mydomain\\myuser%mypasswd plaintext password authentication succeeded error code was NT_STATUS_OK (0x0) challenge/response password authentication succeeded error code was NT_STATUS_OK (0x0)NOTE: both plaintext and challenge/response should return "succeeded." If there is no "challenge/response" status returned then Samba was not built with "--with-winbind-auth-challenge" and cannot support ntlm authentication.
SMBD and Machine Trust Accounts
Samba 3.x
The Samba team has incorporated functionality to change the machine
trust account password in the new "net" command. A simple daily cron
job scheduling "net rpc changetrustpw" is all that is needed,
if anything at all.
Samba 2.2.x
Samba's smbd daemon, while not strictly required by winbindd may be needed to manage the machine's trust account.
Well behaved domain members change the account password on a regular basis. Windows and Samba servers default to changing this password every seven days.
The Samba component responsible for managing the trust account password is smbd. Smbd needs to receive requests to trigger the password change. If the machine will be used for file and print services, then just running smbd to serve routine requests should keep everything happy.
However, in cases where Squid's winbind helpers are the only reason Samba components are running, smbd may sit idle. Indeed, there may be no other reason to run smbd at all.
There are two sample options to change the trust account. Either may be scheduled daily via a cron job to change the trust password.
UglySolution.pl is a sample perl script to load smbd, connect to a Samba share using smbclient, and generate enough dummy activity to trigger smbd's machine trust account password change code.
smbpasswd.diff is a patch to Samba 2.2.5's smbpasswd utility to allow changing the machine account password at will. It is a minimal patch simply exposing a command line interface to an existing Samba function.
Note: This patch has been included in Samba as of 2.2.6pre2.
Once patched, the smbpasswd syntax to change the password is:
smbpasswd -t DOMAIN -r PDC
winbind privileged pipe permissions (Samba-3.X)
ntlm_auth requires access to the privileged winbind pipe in order to function properly. You enable this access by changing group of the winbind_privileged directory to the group you run Squid as (cache_effective_group setting in squid.conf).
chgrp squid /path/to/winbind_privileged
Build/Install Squid
Samba-3.X
As Samba-3.x has it's own authentication helper there is no need to build any of the Squid authentication helpers for use with Samba-3.x (and the helpers provided by Squid won't work if you do). You do however need to enable support for the ntlm scheme if you plan on using this. Also you may want to use the wbinfo_group helper for group lookups
--enable-auth="ntlm,basic" --enable-external-acl-helpers="wbinfo_group"
Samba-2.X<
Squid must be built with the configure options:
--enable-auth="ntlm,basic" --enable-basic-auth-helpers="winbind" --enable-ntlm-auth-helpers="winbind" --enable-external-acl-helpers="wb_group"
Test Squid without auth
Before going further, test basic Squid functionality. Make sure squid is functioning without requiring authorization.
Test the Samba-3.x helper
Testing the winbind ntlm helper is not really possible from the command line, but the winbind basic authenticator can be tested like any other basic helper. Make sure to run the test as your cache_effective_user
# /usr/local/bin/ntlm_auth --helper-protocol=squid-2.5-basic mydomain+myuser mypasswd OKThe helper should return "OK" if given a valid username/password. + is the domain separator set in your smb.conf
Test the Samba-2.2.X helper
Testing the winbind ntlm helper is not really possible from the command line, but the winbind basic authenticator can be tested like any other basic helper:
# /usr/local/squid/libexec/wb_auth -d /wb_auth[65180](wb_basic_auth.c:136): basic winbindd auth helper ... mydomain\myuser mypasswd /wb_auth[65180](wb_basic_auth.c:107): Got 'mydomain\myuser mypasswd' from squid (length: 24). /wb_auth[65180](wb_basic_auth.c:54): winbindd result: 0 /wb_auth[65180](wb_basic_auth.c:57): sending 'OK' to squid OKThe helper should return "OK" if given a valid username/password.
Edit squid.conf
Add the following to enable both the winbind basic and ntlm authenticators. IE will use ntlm and everything else basic:
auth_param ntlm program /usr/local/bin/ntlm_auth --helper-protocol=squid-2.5-ntlmssp auth_param ntlm children 30 auth_param ntlm max_challenge_reuses 0 auth_param ntlm max_challenge_lifetime 2 minutes auth_param basic program /usr/local/bin/ntlm_auth --helper-protocol=squid-2.5-basic auth_param basic children 5 auth_param basic realm Squid proxy-caching web server auth_param basic credentialsttl 2 hours
Note: If your Samba was installed as a binary package ntlm_auth is probably installed as /usr/bin/ntlm_auth, not /usr/local/bin/ntlm_auth. Adjust the paths above accordingly.
Add the following to enable both the winbind basic and ntlm authenticators. IE will use ntlm and everything else basic:
auth_param ntlm program /usr/local/squid/libexec/wb_ntlmauth auth_param ntlm children 5 auth_param ntlm max_challenge_reuses 0 auth_param ntlm max_challenge_lifetime 2 minutes auth_param basic program /usr/local/squid/libexec/wb_auth auth_param basic children 5 auth_param basic realm Squid proxy-caching web server auth_param basic credentialsttl 2 hours
Note: For Samba-3.X the Samba ntlm_auth helper is used instead of the wb_ntlmauth and wb_auth helpers above. This helper supports all Squid versions and both ntlm and basic schemes via the --helper-protocol= option. See the Samba documentation for details.
acl AuthorizedUsers proxy_auth REQUIRED .. http_access allow all AuthorizedUsers
Test Squid with auth
Test browsing through squid with IE. If logged into the domain, a password prompt should NOT pop up.
Confirm the traffic really is being authorized by tailing access.log. The domain\username should be present.
Test with a non-IE browser. A standard password dialog should appear.
Entering the domain should not be required if the user is in the default domain and "winbind use default domain = yes" is set in smb.conf. Otherwise, the username must be entered in "domain+username" format. (where + is the domain separator set in smb.conf)
If no usernames appear in access.log and/or no password dialogs appear in either browser, then the acl/http_access portions of squid.conf are not correct.
Note that when using NTLM authentication, you will see two "TCP_DENIED/407" entries in access.log for every request. This is due to the challenge-response process of NTLM.
References
Joining a Domain in Samba 2.2.x
Authentication is by default disabled in acceleartor mode in Squid-2.X due to conflicts with transparent interception. To enable this feature, at the top of acl.c add the following line:
#define AUTH_ON_ACCELERATION 1
Then "make install".
This feature is somewhat hidden because
The whole concept of "acceleration" in Squid is currently being reworked for the Squid-3.0 release to fix this and a number of other issues.
In Squid, neighbor usually means the same thing as peer. A neighbor cache is one that you have defined with the cache_peer configuration option. Neighbor refers to either a parent or a sibling.
In Harvest 1.4, neighbor referred to what Squid calls a sibling. That is, Harvest had parents and neighbors. For backward compatability, the term neighbor is still accepted in some Squid configuration options.
Regular expressions are patterns that used for matching sequences of characters in text. For more information, see A Tao of Regular Expressions and Newbie's page.
Squid's default configuration file denies all client requests. It is the administrator's responsibility to configure Squid to allow access only to trusted hosts and/or users.
If your proxy allows access from untrusted hosts or users, you can be sure that people will find and abuse your service. Some people will use your proxy to make their browsing anonymous. Others will intentionally use your proxy for transactions that may be illegal (such as credit card fraud). A number of web sites exist simply to provide the world with a list of open-access HTTP proxies. You don't want to end up on this list.
Be sure to carefully design your access control scheme. You should also check it from time to time to make sure that it works as you expect.
SMTP and HTTP are rather similar in design. This, unfortunately, may allow someone to relay an email message through your HTTP proxy. To prevent this, you must make sure that your proxy denies HTTP requests to port 25, the SMTP port.
Squid is configured this way by default. The default squid.conf file lists a small number of trusted ports. See the Safe_ports ACL in squid.conf. Your configuration file should always deny unsafe ports early in the http_access lists:
http_access deny !Safe_ports (additional http_access lines ...)
Do NOT add port 25 to Safe_ports (unless your goal is to end up in the RBL). You may want to make a cron job that regularly verifies that your proxy blocks access to port 25.
$Id: FAQ.sgml,v 1.250 2005/04/22 19:29:50 hno Exp $