If you’re implementing your replication topology with more than one controlling replica, consider using staggered schedules at each of the controlling replicas. If two members of the replica set both attempt to synchronize with a third member at approximately the same time, one of the synchronizations will succeed and the other may encounter a temporary lock failure. The member that attempted — but failed — to synchronize continues to attempt to synchronize, thereby generating additional network traffic and overhead. Using a staggered schedule helps make sure that the target member is not “busy” when the synchronization is initiated.
See Also For more information about synchronizing, see the “Synchronizing Replicas” section later in this chapter.