Problem Summary: When restarting the clusters, the heartbeat is in the Degraded state and the plug-in tcp_udp_plugin is in the No_Response state. On the partner cluster, the process tcp_udp_resp does not exist.
Workaround: Restart the tcp_udp_resp process on the partner cluster by issuing pkill -9 tcp_udp_resp on the partner cluster.