Cluster service fails to start due to fs or clusterfs error 'Cannot mount <device> on <mountpoint>, the device or mount point is already in use!' in a RHEL 6 High Availability cluster with rgmanager

Issue

• When my service fails and tries to recover on node 2, the clusterfs resource won't start:

  Feb 12 00:03:31 node2 rgmanager[6623]: Taking over service service:nfs_export from down member node1.example.com
  Feb 12 00:03:31 node2 rgmanager[16712]: [clusterfs] Cannot mount /dev/dm-13 on /global/archive, the device or mount point is already in use!
  Feb 12 00:03:40 node2 rgmanager[6623]: start on clusterfs "gfs2_archive" returned 2 (invalid argument(s))
  Feb 12 00:03:40 nfs2 rgmanager[6623]: #68: Failed to start service:nfs_export; return value: 1

• An fs resource fails to start with "mount piont is already in use!"