Restoring a Duplicate Master
This topic describes how to recover from a very unusual problem. Do not use the methods described here to resolve the problem unless it is absolutely necessary.
Occasionally the replica that you want to be the master will have a master sequence number that is lower than (or equal to) another master sequence number in the system. When the master detects that its master sequence number is lower than another one in the system, it marks itself as a duplicate master and its process exits. Each time you start the master replica, it will notice that it has been deemed a duplicate master, and its process will again exit.
To force this duplicate' master to become the master and not exit, restart its secd process with the -master_seqno option in the following format:
secd -master_seqno new_master_seqno
where new_master_seqno is a new master sequence number to assign to the replica. Make this number one digit higher than the highest master sequence number in the system.
Use the dcecp registry dump command for each replica to find the highest master sequence number.