Automatic Resolution of the Dual Master Situation

When two stack segments are connected together and no slot number is duplicated on either segment, it is assumed that this is a severed stack rejoin.

It is possible that each stack segment has its own master. Resolution of the dual master situation should generally be in favor of the original stack segment‘s master node. This is because the original stack segment may still retain the unsaved configuration. If the severed segment was restarted before electing a new master node, the unsaved configuration is lost on that segment.

The master election is done using the System UpTime. The master election process collects the System UpTime information of the nodes. If a failover occurs, the System UpTime is inherited by the new master node, and the new master node continues to increase it as time passes. Thus the System UpTime is the time since a master was first elected on a segment. When the stack is broken and both master and backup nodes are on the same segment, the severed segment always has the smaller System UpTime.

If a stack severance results in the master and backup nodes being on different segments, both have the same System UpTime. In this case, the master is elected using the normal node role election method.