Restore Operations ​

The first restore of a database starts by replaying a full backup and any subsequent cumulative or incremental backups that are required to reach the specified restore point. Incremental restores provide an efficient way to continuously update a restored database with incremental changes (based on cumulative or incremental backups). This method is useful for maintaining DR systems and keeping them in synch with source databases. This approach to database recovery is a simpler alternative to setting up database replication. See Running Incremental Restores.

Restore operations provide some flexibility in terms of the selected restore point you want to restore. A restore point is a reference to a particular backup that you want to restore from. You can restore from any point in a backup chain, and you can restore one database while other users continue to access other databases on the same system.

You can specify the restore point for each database in several different ways:

• Use known keywords to specify the latest or oldest backup
• Specify the backup name (the "snapshot")
• Specify a point in time that matches the timestamp for a listed snapshot

Based on the restore point you choose, the system automatically determines the restore path (the sequence of full, cumulative, and incremental backups) that need to be restored, one by one.

Each restore in the sequence is committed separately. If a specific restore fails to commit, the database is restored to the point of the previous successful commit.

Note: During a restore operation, the database that is being restored is accessible to the yellowbrick superuser. It is strongly recommended that you do not log into the database until the restore is complete.

Parent topic:Overview