Sorry folks, false alarm.
This turned out to be a mutation of the LVS ARP problem that resurfaced
after high load was placed on this cluster. I'm going to go sit in the
corner now.
At 10:49 AM 2/20/2002 -0500, you wrote:
I will give it a shot and let you know the results.
At 02:05 PM 2/20/2002 +0100, you wrote:
On Wed, 20 Feb 2002, Alex Kramarov wrote:
> > does TUX work with multiple sockets if you use it on an ordinary
> > (LVS-less) box?
>
> I can assure that tux works on LVS clusters (i use heartbeat from
> www.linux-ha.org) - and it works perfectly, even doesn't need any
> restart/hup/other during custer failover. i simply do net setup it to
> listen on different ip's , on my machine it simply listens on 0.0.0.0
yes, perhaps the specific listen on IP addresses is doing something
strange on LVS. Mike, could you try Alex's suggestion and bind all TUX
threads to the 0.0.0.0 listen address? Does that fix your setup?
Ingo
_______________________________________________
tux-list@redhat.com
https://listman.redhat.com/mailman/listinfo/tux-list
_______________________________________________
tux-list@redhat.com
https://listman.redhat.com/mailman/listinfo/tux-list