We have created an active active pacemaker cluster for Filesystem which is being used by Liberty application (IBM).
Whenever we are testing a failover the switch is happening perfectly but when the one node comes back instead of bringing the FS resource on that node the cluster is restarting the resource and is setting up fresh on both nodes.
This is making the application services to restart which is impacting the application.
Generally, when the node which is down comes up the FS resource should come up on that node without hindering the active resource on other node.
Is this a normal behaviour of pacemaker or not.
Any help or suggestions are welcome and will be helpful.