Hello.
I'm not sure, but maybe your slave node is in the 1st place in the failover order, so it tries to take over the service. If the other ser isn't releasing the service, maybe there is also some problem with the script that stops the service and releases the resources. Have you tried to do a service takeover between the nodes to check if it works ok?
I was always advised to not put the cluster services on the server startup process, to avoid this situation, in any kind of OS and cluster service (MS, IBM HACMP, RedHat Cluster, etc...), so I start the cluster daemon only after verifying if the machine is ok.
|