Example of clustered environment with three Domino servers
The environment is a Domino clustered environment with three Domino servers as members of the Domino cluster. Server C is identified as your backup server and is running archive-style transaction logging. Servers A and B are running circular-style or linear-style logging. To provide high availability and load balancing of mail, replicas exist on several servers. Databases A-L exist on servers A and C. Databases M-Z exist on servers B and C. A successful full backup of all databases was completed earlier in the week. Successful incremental backups of the archive-style transaction logs are completed every four hours, with the last one completing two hours ago. Today a user discovers that 30 mail messages were accidentally deleted. They were deleted yesterday at about 3:30 P.M. and were stored in the database mander.nsf.
To recover accidentally deleted mail messages do the following:
On server C, perform a point-in-time recovery of the following database:
UNIX or Linux: /mail/mander.nsf
Windows: mail\mander.nsf
Select the database from the last successful backup of the database (for example, the full backup that was completed successfully earlier in the week). Begin the restore.
On the Lotus Notes tab of the Restore Marked Files dialog box, select the option. Select the option. Specify yesterday's date at 3:25 P.M. (the time right before the user deleted the mail messages) as the point-in-time for recovery.
A version of the database is recovered to yesterday at 3:25 P.M. and a database that contains the deleted messages should exist on server C. Verify the existence of messages in the mander.nsf database on server C. If all is as expected, copy the accidentally deleted messages from the database on server C to the database on server B.
After the copy, verify the existence of the messages in the database on server B. If all is as expected, from server B create a new replica of the following database on server C.
Windows: mail\mander.nsf
UNIX and Linux: /mail/mander.nsf
Cluster replication should now be functional for the database on servers B and C.
In this example the option is chosen. If the option was chosen instead, the final results of the recovery would be different. The restored database would function the same, except that the replica ID would not be changed. The database would be recovered to the specified point-in-time. However, because the replica ID was not changed during the restore, it would match the replica ID on server B. Therefore, all the changes to the database between the point-in-time to which the database was recovered and the current time (including the deletion of the 30 mail messages) eventually is replicated to the database on server C. Two identical copies of the database exist, one on server B and the other on server C. Both copies are the same as when the restore was started on server C.