Backup and restore

Because the index properties on Image Services and Content Engine are synchronized, the two systems must be backed up at the same time. If the system backups are not synchronized, resynchronization after a restore could become more complicated. Choose a time to make your backups when

When either system needs to be restored, the other system must also be restored to the same level.

The first line of defense is mirroring or RAID. We recommend that you always use mirroring or RAID. The last line of defense is backup/restore.

To back up a large system, we recommend mirroring the database server.

To perform a backup

  1. Break the connection to one mirrored server, leaving the remaining mirrors in production.
  2. Isolate the one server and back it up.
  3. When the backup is finished, re-establish the mirror connection, and the servers will automatically resynchronize.

This method avoids bringing the system down for backup and prevents a degradation of performance on the production system during the backup. What is in effect being backed up is the state of the system as it would be if the system had crashed at the time the mirror was broken.

For Image Services systems, you can use FileNet’s EBR (Enterprise Backup and Restore) to back up the system. The backup should be performed while both the Image Services system and the Content Engine system are shut down.

When both the Image Services system and the Content Engine system are backed up at the same time, either system can be restored first. Both systems should be restarted at about the same time.

Synchronizing the CE and IS Catalogs

In general, when a database is restored, the recovery log of the database can be used to roll the database forward in time so that no completed database transactions are lost. The ideal situation after a restore is that the database is rolled forward to the last complete transaction. Then, no work is lost, and production can be resumed.

The roll forward technique does not apply to databases such as the Image Services transient database, however, because the transient database maps data stored in files or partitions, and those files or partitions cannot be rolled forward. When such databases are restored, the files or partitions that they map must also be restored, and updates made since the backup are lost. David?

Fortunately, the Content Engine catalog database and the Image Services catalog database are databases to which the roll forward technique applies. The best practice is to take advantage of database roll forward.

If mirroring or RAID and roll forward have both failed, you could restore the other catalog database as well. If you have followed best practices and backed them both up at the same time, the two catalog databases will be synchronized. However, the work since the backups were made will be missing from both systems.
It is desirable that the restore of one catalog database does not force a restore of the other. In this regard, several cases must be considered:

For documents entered on the Image Services system, dual catalog entries must exist until the Panagon applications are switched over completely to the FileNet P8 Platform applications and cataloging to the Image Services catalog is no longer performed. When Image Services cataloging is turned off, there is no catalog synchronization issue.