About Exchange Instant Recovery methods
If the snapshot is preserved with the Instant Recovery option, NetBackup restores the database using rollback of the snapshot volume(s) when appropriate. Usually, a rollback of the snapshot volume(s) that contain the Exchange files is the fastest way. However, the whether or not a rollback is appropriate depends on several things: the configuration of the Exchange database files, the contents of the volumes, and configuration of the disk array. If a volume rollback cannot be performed, the files that are required for restoration are copied from the snapshot volume to the destination volume. Instant recovery of Exchange differs from Instant Recovery of a file system. For Exchange, NetBackup decides which recovery method to use. For file system restores, the user chooses the Instant Recovery method.
NetBackup uses the following methods during an Exchange database restore to restore the physical files:
Volume rollback | The entire volume is rolled back, or resynchronized, by using the snapshot. This method overwrites the entire volume with the snapped volume. |
File copy back | Individual files are copied back from the snapped volume to the current volume. |
To determine if a volume can be rolled back, checks are made to insure that the same list of files exists in the following places:
The snapshot volume is compared with the cataloged list of files to restore. These lists must match exactly. An example of a difference is a file that was included in the snapshot, but was not cataloged because it is not an Exchange file. The snapshot is not rolled back because that action overwrites the non-Exchange file. Exchange files also may exist on the snapshot but not in the catalog if the backup did not include all the databases on the volume.
The snapshot volume is compared with the current volume. All files on the current volume must also exist in the snapshot. If there is a file that is not on the snapshot, a rollback is not performed because that action does not restore that file.
In both comparisons, NetBackup excludes certain files from consideration. For example, unneeded Exchange transaction logs, files Exchange re-generates, or any files that are artifacts of the NetBackup process. The bppfi log shows when such a file difference is found and excluded from consideration.
The copy-back restore method is used in the following situations:
If the system provider is used and the snapshot selected for restore is not the most recent snapshot
If there are other files on the volume that could be lost
If all the files on the snapshot are not selected for restore
If you select . The copy-back method must be used for the volume that contains the log files. A roll-forward recovery needs the log files that were created since the backup. A rollback cannot be performed since it removes those log files. If the database file (
.edb) is on a different volume, that volume is still evaluated with the other criteria to determine if it is eligible for rollback.
If multiple volumes are included in the restore set, each volume is evaluated separately to determine if it is eligible for rollback. (The restore set is based on the location of the Exchange database, transaction logs, and system files that are part of the restore.) For example, perhaps the volume that contains the database files is eligible for rollback, but the volume that contains log files has extra, non-Exchange files. At the time of the restore, only the volume that contains the database files is rolled back. All the log files are copied back from the snapshot to the current volume.