rgmanager "Shutting down uncleanly" in RHEL or RHCS

Issue

• rgmanager died unexpectedly and could only rejoin the cluster after a reboot of the machine, in RHEL 4:

May  6 09:08:31 node02 clurgmgrd[5447]: <warning> #67: Shutting down uncleanly
May  6 09:08:31 node02 ccsd[7947]: Cluster manager shutdown.  Attemping to reconnect...
May  6 09:08:31 node02 kernel: CMAN: Being told to leave the cluster by node 1
May  6 09:08:31 node02 kernel: CMAN: we are leaving the cluster.
May  6 09:08:31 node02 kernel: SM: 01000005 sm_stop: SG still joined
May  6 09:08:31 node02 kernel: SM: 03000003 sm_stop: SG still joined
May  6 09:08:31 node02 ccsd[7947]: Cluster is not quorate.  Refusing connection.
May  6 09:08:31 node02 ccsd[7947]: Error while processing connect: Connection refused
May  6 09:08:31 node02 ccsd[7947]: Cluster is not quorate.  Refusing connection.
May  6 09:08:31 node02 ccsd[7947]: Error while processing connect: Connection refused
May  6 09:08:32 node02 ccsd[7947]: Cluster is not quorate.  Refusing connection.
May  6 09:08:32 node02 ccsd[7947]: Error while processing connect: Connection refused
May  6 09:08:32 node02 ccsd[7947]: Cluster is not quorate.  Refusing connection.
May  6 09:08:32 node02 ccsd[7947]: Error while processing connect: Connection refused
May  6 09:08:42 node02 ccsd[7947]: Unable to connect to cluster infrastructure after 6960 seconds.

• In a RHEL 6 cluster, when a node is evicted or removed from the cluster, rgmanager shuts down uncleanly but never competely exits:

Apr 27 09:52:13 rgmanager #67: Shutting down uncleanly