nfsserver resource starts on multiple nodes in pacemaker cluster when node re-joins cluster or command `pcs resource cleanup` was run

Issue

When one of the cluster nodes is gracefully rebooted and it re-joins cluster the nfsserver resource gets started on rebooted node despite fact that it is already running on other node.

# pcs status 
.
     nfs-daemon (ocf::heartbeat:nfsserver):     Started[ node 1 node2]  
.

Once pacemaker spots that resource (which is not clone resource) is running on more locations it triggers corrective mechanism and stops nfsserver resource on all nodes. Then resource starts on one node only. This however brings service interruption which is not necessary in this situation:

Aug 7 11:25:56 node2 pengine[2461]:   error: Resource nfs-daemon (ocf::nfsserver) is active on 2 nodes attempting recovery

Similar behaviour happens when pcs resource cleanup is run (nfsserve resource is started on all cluster nodes)