Re: [squid-users] High availability based on Squid process

From: Chris Robertson <>
Date: Fri, 25 Apr 2008 14:36:49 -0800

J. Peng wrote:
> On Wed, Apr 16, 2008 at 2:09 AM, BJ Tiemessen <> wrote:
> Linux LVS works very well for this situtation.
> We once had 200 squid hosts and use 2 LVS hosts before them.
> LVS run with DR or Tunneling mode.

Was this in a forward proxy or an acceleration setup?

If it was a forward set up, given that a lot of SSL sites don't like the
requesting IP to change (and I assume in anything but a LVS-NAT setup
the individual Squids make connections from individual real IP
addresses), how did you go about "encouraging" a client's connections
towards a single back end? From what I've read, the load balancing
methods focus on round-robin and least-load. Or do you just SNAT the
requests Squid makes?

Received on Fri Apr 25 2008 - 22:37:03 MDT

This archive was generated by hypermail 2.2.0 : Thu May 01 2008 - 12:00:04 MDT