From TN 1228354. This is the last option
- Ensure that the corrupt replica is not in use by any user or server.
On the Domino server hosting the corrupted database, enter the following command: “dbcache flush” (command may require entering two or three time to release the database). - Using operating system commands, delete the corrupt replica file.
- Using the Lotus Notes client, open the “good” replica located on the other clustered server.
- Create a new replica on the server where the “bad” replica was deleted.
Select File -> Replication -> New Replica. - Complete the replication.
- Stop the Cluster Directory task on the server where the new replica was created by running the console command “tell CLDBDIR quit”
- Restart Cluster Directory task by running the command “load CLDBDIR”
- Verify database replica is on clustered server: Open Cluster Directory database and select the view, Databases by Replica ID
- Verify that each replica copy of the same NSF file has the same Replica ID value.