Dear all:

There are two nodes cluster service on OES Linux SP1,
(The same hardware and software environment on the two nodes)
and connected to san storage correctly.

What happening when the ethernet failed?

If I shutdown one of the node which have resource on it,
then it will failover to another one. it looks good.

But if I pull out the cable from the ethernet (to simulate the ethernet failed),
I discover the node which have resource on it will reboot another node,
the resource node was not failover to the alive node, but reboot the other node?
and the resource node was still disconnected to the ethernet.

After the other node (which was be reboot) start up, the cluster service still failed,
the resource still on the wrong node (which was already disconnected to ethernet).
I don't understand why? I think it is abnormal, isn't it?

The above condition on Netware is no problem on it, whether or not
I shutdown or I pull out the cable from ethernet on the resource node,
it will failover to another node correctly.

Please help me, I tried to find an answer to resolve the problem on Linux cluster service,
how can it failover to another correctly when I pull out cable from the ethernet?