qmgr service restarts multiple times during cluster failover on RHCS

Issue

While testing local failovers between nodes, every time a failover was simulated (by crashing an active server), the queue managers that failed over to the standby server restarted a couple of times on the standby server before settling in the “Running” state. Sometimes, they failed to settle in the Running state. A side-effect of the queue manager restarting multiple times was that there were multiple candle instances that became active for 1 queue manager (instead of 1 candle process per queue manager).