I'm trying to set up a small cluster using the High Availability and
High Availability Management groups. I'd like to run this on 3 different
machines, one being a "management" machine and the other two being the
actual cluster. It doesn't seem to want to cooperate. I can set the
cluster nodes up just fine, working from the management server. As soon
as I try adding an "IP address" to the cluster, everything goes south. I
can no longer use luci on any of the three machines.
I can start luci on any of the 3 servers, but the cluster doesn't show
any more. If I try and recreate the cluster, it indicates the cluster
already exists. I can't find anything in the logs other than a bunch of
python errors similar to perl errors, and those don't really seem to be
indicative of hard errors.
I've got this scheme running successfully on 3 virtual machines on a
single host, and that seems to work just fine. Real servers are beating
me up.
Any suggestions, please?
steve campbell