Re: SV: Hot Standby

Mattias Paulsson wrote:
> The last couple of days I've been researching ways to solve a much similar problem and I've come up with the following URLs that might be of help:
>
> Linux HA Project: http://www.henge.com/~alanr/ha/
> EddieWare http://www.eddieware.com
> Fake: Redundant Server Switch http://linux.zipworld.com.au/fake/
>
> Within a couple of days I will have tested these methods but until then they're just pointers that might be useful.
>
> There is probably expensive hardware that will do the trick but let's try a simpler approach, right?

Maybe my reply didn't get through. Current developmental versions of Squid
support Cisco WCCP 1.0. WCCP is a very simple protocol which provides high
availability and autodetection of cache servers. This is how it works:

On your Cisco, enable WCCP (11.1 or later):

rtr# conf t
rtr(config)# ip wccp enable
rtr(config)# int e0
rtr(config-if)# ip web-cache redirect
rtr(config-if)# end
rtr# copy running-config startup-config

With Squid 2.3, add the following to your config file:

wccp_router your.routers.ip.address

Squid should automatically start sending keepalive messages to the cisco
router. The Cisco router will proxy requests over to your Squid server as long
as Squid keeps sending keep alives. If Squid dies or the machine crashes, the
router will use another WCCP enabled cache or route out normally.

If you have more than one cache on your network, the Cisco should load balance
(I believe, it says it supports multiple caches, but I'm not sure how it works
exactly :).

Now obviously, this has it's problems. WCCP 1.0 is suseptable to certain,
unpleasent security problems... there is no authentication scheme so you will
probably want to put in some firewall rules. WCCP 2.0 fixed this problem, but
it's currently not implemented in Squid (as far as I know).

Jordan

--
Jordan Mendelson     : http://jordy.wserv.com
Web Services, Inc.   : http://www.wserv.com