if the secondary failover peer does not get the "split" option in its configuration, how does it know if it is responsible for a given client?