[Linux-cluster] Unwanted cluster automigration after first node joins cluster

Miroslav Zubcic mvz at nimium.hr
Mon Jul 11 16:55:02 UTC 2005

I have another problem with Red Hat Cluster Suite:

Let's say one (active) node in 2-node cluster failed. It is rebooted
fenced etc... and services are migrated on the second node. That's OK.

But when I repair and boot again first (previously failed) node and
start cluster daemons with /etc/init.d/clumanager, services are
MIGRATED BACK again (!!!) automaticly, couple of seconds after firing
up first node.

Since I have equal hardware for all nodes in cluster, and REALLY
critical services, I'm experiencing unneeded service failover.

How can I configure RH CS *NOT* to failover services back on the first
node after first node is up again?


Miroslav Zubcic, RHCE, Nimium d.o.o., email: <mvz at nimium.hr>
Tel: +385 01 4852 639, Fax: +385 01 4852 640, Mobile: +385 098 942 8672
Mrazoviceva 12, 10000 Zagreb, Hrvatska

More information about the Linux-cluster mailing list