Restarting a Restore

Enterprise Edition Home | Express Edition Home | Previous Page | Next Page Overview of the ON-Bar Backup and Restore System > Restoring Data with ON-Bar > Using Restartable Restore to Recover Data (IDS) >

Restarting a Restore

You can restart a point-in-time, whole-system, or parallel restore. The physical restore restarts at the storage space and level where the failure occurred. If the restore failed while some, but not all, chunks of a storage space were restored, even a restarted restore must restore all chunks of that storage space again. If storage spaces and incremental backups are restored successfully before the failure, they are not restored again.

Figure 14 shows how a restartable restore works when the restore failed during a physical restore of dbspace2. For example, you set RESTARTABLE_RESTORE to ON before you begin the restore. The level-0, level-1, and level-2 backups of rootdbs, and the level-0 and level-1 backups of dbspace1 and dbspace2 are successfully restored. The database server fails while restoring the level-1 backup of dbspace2. When you restart the restore, ON–Bar restores the level-2 backup of dbspace 1, the level-1 and level-2 backups of dbspace2, and the logical logs.

Figure 14. Restartable Physical Restore

begin figure description - This figure is described in the surrounding text. - end figure description

Interaction Between Restartable Restore and BAR_RETRY Value

If BAR_RETRY > 1, ON–Bar automatically retries the failed storage space or logical log. If this retry is successful, the restore continues and no restart is needed.

If BAR_RETRY = 0 or 1, ON–Bar does not retry the failed storage space or logical log. If the database server is still running, ON–Bar skips the failed storage space and attempts to restore the remaining storage spaces.

Table 13 shows what to expect with different values for BAR_RETRY in a different restarted restore example.

Table 13. Restartable Restore Results with Different BAR_RETRY Values
ON–Bar Command	BAR_RETRY = 2	BAR_RETRY = 0
onbar -r dbs1 dbs2 dbs3	restore level-0 dbs1, dbs2, dbs3 restore level-1 dbs1 FAILS restore level-1 dbs1 RETRY PASSES restore level-1 dbs2, dbs3 restore level-2 dbs1, dbs2, dbs3 restore logical logs	restore level-0 dbs1, dbs2, dbs3 restore level-1 dbs1 FAILS
onbar -RESTART	No restart is needed because everything was successfully restored.	restore level-1 dbs1, dbs2, dbs3 restore level-2 dbs1, dbs2, dbs3 restore logical logs
onbar -r dbs1 dbs2 dbs3	restore level-0 dbs1, dbs2, dbs3 restore level-1 dbs1 FAILS restore level-1 dbs1 RETRY FAILS restore level-1 dbs2, dbs3 restore level-2 dbs2, dbs3 restore logical logs onbar -r dbs1 dbs2 restore level-1 dbs1 restore level-2 dbs1 restore logical logs	restore level-0 dbs1, dbs2, dbs3 restore level-1 dbs1 FAILS onbar -RESTART restore level-1 dbs1, dbs2, dbs3 restore level-2 dbs1, dbs2, dbs3 restore logical logs

Restarting a Logical Restore

If a restore fails during the logical phase and you restart the restore, ON–Bar verifies that the storage spaces have been restored successfully, skips the physical restore, and restarts the logical restore. Figure 15 shows a cold restore that failed while restoring logical log LL-3. When you restart the cold logical restore, log replay starts from the last restored checkpoint. In this example, the last checkpoint is in logical log LL-2.

If a failure occurs during a cold logical restore, ON–Bar restarts it at the place that it failed.

Important:

If a failure occurs during a warm logical restore, you have to restart it from the beginning. If the database server is still running, use the onbar -r -l command to complete the restore.

Figure 15. Restartable Cold Logical Restore

Set the RESTARTABLE_RESTORE parameter to ON. A restartable restore makes the logical restore run more slowly if many logical logs need to be restored, but it saves you time if something goes wrong and you need to restart. Restartable restore does not affect the speed of the physical restore.