Repeated membership changes from corosync in RHEL 7 Pacemaker Cluster resulting in OOM errors on the nodes

Issue

Corosync is repeatedly experiencing membership updates on one or more nodes, flooding the logs with the following message:

corosync[pid]:  [TOTEM ] A new membership (<ip addr>:1844600) was formed. Members
corosync[pid]:  [QUORUM] Members[8]: 6 7 1 2 3 4 5 8
corosync[pid]:  [MAIN  ] Completed service synchronization, ready to provide service.

This coincides with slowly increasing memory consumption eventually resulting in Out of Memory errors in the environment.