The memory footprint of "corosync" process increases when it has communication problems with the other cluster Nodes

This document (7022840) is provided subject to the disclaimer at the end of this document.

Environment

SUSE Linux Enterprise Server 12 Service Pack 2 (SLES 12 SP2)

Situation

When corosync experiences communication problems with the other cluster Nodes, the "corosync" process memory footprint started increasing until out of memory occur.

Resolution

This memory leak bug is fixed in packages "corosync" and "libcorosync4" version 2.3.5-6.23.1 and newer, for SUSE Linux Enterprise High Availability Extension 12 SP2.

Cause

If the connection between the Nodes in a HA cluster is interrupted short enough to break the Node connection but not long enough to trigger a fence, the memory footprint of the corosync process will grow. This can lead to out of memory error if this condition exist long enough.

Additional Information

Everytime corosync prints the following type of error messages, the memory footprint increases:
2018-02-27T13:52:10.477274+11:00 Node1 corosync[42496]: [TOTEM ] A processor failed, forming new configuration.
2018-02-27T13:52:12.549516+11:00 Node1 corosync[42496]: [TOTEM ] A new membership (10.0.9.13:75200) was formed. Members left: 2
2018-02-27T13:52:12.549919+11:00 Node1 corosync[42496]: [TOTEM ] Failed to receive the leave message. failed: 2

Disclaimer

This Support Knowledgebase provides a valuable tool for SUSE customers and parties interested in our products and solutions to acquire information, ideas and learn from one another. Materials are provided for informational, personal or non-commercial use within your organization and are presented "AS IS" WITHOUT WARRANTY OF ANY KIND.