This issue typically comes about when a Veritas Global Cluster has been failed over improperly, or if the Replication Network option of Takeover was manually used to move replication from one site to another. This is because the Replication Network has been told that the Primary site no longer exists and the Secondary is now to become the permanent Primary site.

Generally, a graceful failover of the cluster, or choosing the 'Outage' option instead of Takeover will prevent this from happening.

One resolution to this issue is to delete and re-create the Replicated Data Set (RDS) that hosts both the Primary and Secondary sites. This will require a full resync of the volume data; however, any resolution at this point would more than likely require a full resync as well.

Perform these steps to delete / re-create the RDS:

1) Write down the following items from the current RDS configuration:

- Hostnames for each of the hosts at both the Secondary and Primary sites

- IP addresses being used for Primary and Secondary sites

- The name of the RDS and the name of the RVG (Replicated Volume Group)

2) In Replication Network, right click on the current RDS and select the DELETE option

3) Right click again on the host name and select CREATE REPLICATED DATA SET

4) A wizard will come up and provide a walk through on the process of creating the new RDS. While going through the wizard, it will be required to enter the information gathered in Step 1.

5) Once the wizard completes it, the new RDS (with the same name as the original) should be seen in the Replication Network GUI. 

Right click on the RDS and select Add Secondary in order to add the Secondary replication site.

6) Once both the Primary and Secondary sites are listed in the new RDS, right click on the Primary and choose START REPLICATION.

At this stage, replication should begin a complete resync of data, and continue un-interrupted.