Re: FW: Multiple cache servers with failover ?

From: Martin Hamilton <[email protected]>
Date: Wed, 15 Mar 2000 11:06:20 +0000

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Content-Type: text/plain; charset=us-ascii

"Carpenter, Dean" writes:

| I sent this in last week, and haven't heard a peep. Is there no one out
| there using (or trying to) squid in a failover/HA environment ?

Hi - sorry for the late reply!

In the UK we (the JANET Web Cache Service [1]) are using Linux Virtual
Server [2] to load balance two (soon to be three) clusters of quite
busy Squids. Overall our server farms are shipping some 70 million
proxy HTTP requests/day at peak loadings, but this isn't entirely load
balanced using LVS yet.

At the moment the code we use for load balancing and monitoring is
entirely homegrown, but it's likely that we'll move to RedHat's
Piranha [3] clustering (new in RedHat 6.2 beta [4]). At the moment
we're using IP tunnelling rather than NAT, and server status
monitoring is initiated by the real server rather than the director,
so our model isn't quite compatible with what Piranha is doing - it
only supports NAT, and has the director polling the real servers.

Time to pull the hacking gloves on... :-)

Cheers,

Martin

[1] <URL:http://wwwcache.ja.net/>
[2] <URL:http://www.linuxvirtualserver.org/>
[3] <URL:http://www.redhat.com/support/wpapers/piranha/>
[4] <URL:ftp://ftp.redhat.com/pub/redhat/redhat-6.2beta>

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.1 (GNU/Linux)
Comment: Exmh version 2.1.1 10/15/1999 + martin

iD8DBQE4z26sVw+hz3xBJfQRAgjSAJ0aHRg2ekVgLgTH/Lvrf0OeMYXQXgCeIPfw
KvnX4kdcaQCfuYJ/5nrn+74=
=FUkS
-----END PGP SIGNATURE-----
Received on Wed Mar 15 2000 - 04:11:54 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:52:08 MST