corosync is repeatedly reporting "A processor joined or left the membership" every few seconds and corosync is using a lot of memory in a RHEL 6 or 7 High Availability cluster

Issue

• I see corosync reporting over and over every few seconds that "A processor joined or left the membership"

Jan 15 12:42:53 node1 corosync[1804]:   [TOTEM ] A processor joined or left the membership and a new membership was formed.
Jan 15 12:42:53 node1 corosync[1804]:   [CPG   ] chosen downlist: sender r(0) ip(10.10.10.1) ; members(old:2 left:0)
Jan 15 12:42:53 node1 corosync[1804]:   [MAIN  ] Completed service synchronization, ready to provide service.
Jan 15 12:42:56 node1 corosync[1804]:   [TOTEM ] A processor joined or left the membership and a new membership was formed.
Jan 15 12:42:56 node1 corosync[1804]:   [CPG   ] chosen downlist: sender r(0) ip(10.10.10.1) ; members(old:2 left:0)
Jan 15 12:42:56 node1 corosync[1804]:   [MAIN  ] Completed service synchronization, ready to provide service.
Jan 15 12:43:00 node1 corosync[1804]:   [TOTEM ] A processor joined or left the membership and a new membership was formed.
Jan 15 12:43:00 node1 corosync[1804]:   [CPG   ] chosen downlist: sender r(0) ip(10.10.10.1) ; members(old:2 left:0)
Jan 15 12:43:00 node1 corosync[1804]:   [MAIN  ] Completed service synchronization, ready to provide service.

• corosync seems to be using a lot of memory and there are repeated "A processor joined or left the membership" messages in the logs
• corosync caused an oom-killer kill after the node had been in the cluster for some time.