[squid-users] SQUID - Best Practices

I'm looking for a proven set of "configure" options to build SQUID
2.5-Stable 10 and a SQUID.CONF file that will give the best performance
while conserving the bandwidth on our T-1's. Our network topology is 16
spokes connected to a hub site by Point-To-Point T1's. The network is home
to 2,000 Windows desktops, 700 Windows laptops and 50 Windows servers.
System wide, the web traffic is a 40/60 ratio static and dynamic pages.

We have 2 Squid servers in a round-robin configuration at the hub site
connecting to our ISP and 12 of the spoke sites have a "child" Squid
running. At the hub site our server hardware is dual 800MHz P3 CPU's, 1GB
RAM, 40GB IDE disk, 100MB/s switched Ethernet. The software platform is
SuSE 9.2 Pro (no GUI), Squid 2.5 STABLE 9 and SuSE Firewall. At 12 of the
spoke sites, our server hardware platform is single 1GHz P3 CPU, 512MB RAM,
20GB IDE disk. The software platform is a minimal SuSE Pro 9.3 (no GUI)
with Apache 2.0.54 and Squint 0.3.16 for log monitoring.

At each of the spoke sites, Squid was built from
Squid-2.5.STABLE10-20050725.tar.gz using configure options:
--enable-dmalloc
--enable-storeio=aufs,ufs
--enable-carp
--enable-icmp
--enable-delay-pools
--enable-snmp
--enable-err-languages=English
--enable-default-err-language=English
--disable-hostname-checks
--enable-underscores
--enable-stacktrace.

The SQUID.CONF is basically the default settings with following exceptions:
Cache-Dir is aufs size=1024 TopLevelDirs=16 and SecondLevelDirs=256
Cache_peer (Parent IP) parent 80 3130 round-robin

Anyone have any good suggestions?

Thanks in advance.

Bob Morrison, CNE, MCSE
Network Administrator
Wallingford CT Public Schools
Received on Thu Jul 28 2005 - 20:51:40 MDT