Why do the IPaddr2 resource failback to a node even resource-stickiness property is set ?

Issue

• While triggering a manual fence of active node using pcs stonith fence command, node gets rebooted and VIP resources failover to a passive node. But the same VIP resource failback to the node which was fenced at first place, even though resource-stickiness property is configured under cluster.
• When a fenced node joins back the cluster membership, we can find that for a moment the IPaddr/IPaddr2 resource shows started on two nodes and then starts on the node which joined the cluster membership after getting fenced.
• The cluster property resource-stickiness fails to make the VIP resource stick to the node and it fails back to the node which joined back the cluster membership after getting fenced.