I currently have a 2 node OES Linux cluster (which we will be upgrading to
3 nodes) with 4 virtual servers. Everything seems to be working fine.
I'm just trying to verify everything is configured optimally. First one
minor thing. If I reboot one of the nodes that has resources currently on
it they all fail over. But in the log it states that the resource is
unassigned then it goes to the other node. Is this normal? I have a
priority set for all the resources that include both servers.

Another question is I had as situation happened recently that maybe
someone can shed some light on. I needed to reboot one of the cluster
nodes, when I did the resources tried to fail to the other server and
couldn't, the SBD did but not my virtual servers, they went comatose. I
then got the server back up and tried to migrate the resources back, they
were still comatose. I tried to offline and online them but they wouldn't
work. It turns out that the second server (that I originally tried to
fail to) had some bad RAM. The only way I could get the resources back
online is to shut down the server with bad RAM and then reboot the other
node. I then pulled the bad RAM started the server and it joined up fine
(albiet 2gb RAM lighter). My question is, and forgive my inexperience,
what exactly causes a resource to go comatose. Also how would a cluster
expert handle this situation I couldn't help but feel that someone with
more experience would have handled it differently (other that rebooting
the nodes till the resources came back online).