On Fri, 17 Sep 2004 11:55:45 -0500 (CDT), Adam Maloney <adam at whee.org> wrote:
> Dan's message reminded me to ask for real-world implementations of large,
> highly-available squid installations...on Linux (just so it's on-topic).

I tried to use Squid about a year ago as a reverse proxy.  We had a
lot of free hosting content and needed to distribute the disk
transactions across a number of disks.  There was too much content to
simply replicate it across machines, so the idea was to use Squid to
dynamically replicate the content as it was requested.  It worked fine
in testing, but under load everything slowed to a crawl.

I spent about a week tweaking it and never could get it to work right.
 Squid tries to do everything and requires a very complex
configuration file that isn't well documented.  Figuring out how to
get it to perform better requires guesswork or reading the source.

In the end I gave up and spent a few hours writing an ISAPI filter for
Zeus that solved the problem in a much better way than squid ever
could.  Everything I read indicated that no one was using Squid with
that much traffic (50-100mbits per machine or more).  Maybe you'll
have better luck using it as an outgoing proxy, but someone would have
to try very hard to convince me to look at Squid again.

-- 
David Phillips <david at acz.org>
http://david.acz.org/

_______________________________________________
TCLUG Mailing List - Minneapolis/St. Paul, Minnesota
Help beta test TCLUG's potential new home: http://plone.mn-linux.org
Got pictures for TCLUG? Beta test http://plone.mn-linux.org/gallery
tclug-list at mn-linux.org
https://mailman.real-time.com/mailman/listinfo/tclug-list