Currently, by design, the cluster probe needs the hostname and node name to be the same in the cluster environment. If the hostname is changed in the middle of monitoring, the cluster probe will break and give the following error in the logs:?
Mar 20 17:42:07:502 cluster: SetupNodes - node=rhel-cluster-node1.test.com address=192.168.2.100?
Mar 20 17:42:07:502 cluster: SetupNodes - node=rhel-cluster-node2.test.com address=192.168.2.101?
Mar 20 17:42:07:502 cluster: InitalizeCluster - this node 'rhel-cluster-node1.test_p.com' was not found in the configuration. Trying the shortname 'rhel-cluster-node1'?
Mar 20 17:42:07:502 cluster: InitalizeCluster - this node 'rhel-cluster-node1' was not found in the configuration?
Mar 20 17:42:07:502 cluster: Initialize Cluster failed!?
Another symptom is the cluster probe does not allow the user to create a profile or shared section. It will throw an error ?The node has no valid cluster probe address and you can?t add any resources until this is resolved?.?
As of March 2013, there is an Idea posted at the following URL to have the probe work independently irrespective of the hostname/nodename. Please vote (promote) the Idea.
In the meantime, the workaround is to manually make the change in the cluster probe configuration (Raw Configure - SHIFT+RightClick). In the setup folder, adjust the "Node" key value to match what is in the cluster.conf file in order to make the probe work again.?
If the node key does not exist, manually create it.