[Linux-cluster] Problem with fenced on 2 node cluster

[Lon Hohberger]

On Tue, 2008-01-29 at 12:05 -0600, Brad Filipek wrote:
> I have a 2 node cluster setup fairly basic right now (mounts an ext3
> filesystem, and activates a virtual IP). When both boxes are powered
> off, I can start up node1 just fine. It starts the services and runs
> great. Then I power up node2 and it joins fine. However, if I run "ps
> aux | grep fenced" on node1 I see that /sbin/fenced is running and it
> never goes away. Is there anyway to check and see what it is doing? I
> know that the fenced service works because it will reboot the other
> node if I disconnect the network cable. I am using an APC master
> switch for power. 

fenced should stay running at all times... did you mean it's stuck in an
infinite loop?

fenced spawns agents - such as 'fence_apc' - which should only be
present during a fencing operation.  If those stay around for a long
time, there's a problem.

-- Lon