Restoring an Oracle Database to a Point-in-Time

Use a point-in-time restore to revert the database to a state before an undesired transaction or before a point of failure.

Restore the control files when you restore and recover an entire database to a previous point-in-time from an online backup or offline backup to the original host.

Before You Begin

  1. Perform an online full backup if one is not available.
  2. Clear the Oracle Flashback area which forces RMAN to use the SnapProtect objects when performing the restore.
  3. Set the database to the correct mode defined in the table, or select the Switch Database mode for Restore check box (Options tab), to set the SnapProtect software to automatically move the database to the correct mode before the restore.
    Control File in the Restore Required Database Mode
    Control file is included in the restore NOMOUNT
    Control file not included in the restore MOUNT
  4. Optional: When the backup pieces are unknown to the RMAN repository, catalog the backup pieces. For more information, see Cataloging RMAN Backup Pieces.

Procedure

  1. From the CommCell Browser, expand Client Computers > client > Oracle.
  2. Right-click the instance, point to All Tasks and then click Browse and Restore.
  3. Select Latest Backup and click Time Range.
  4. Select the Absolute Time option and enter the End Time date and time.

    For Oracle agents the default Time Zone is the client's time zone. For all other agents the default Time Zone is the CommServe computer's time zone.

  5. On the Browse view, select the data to restore in the right pane and click Recover All Selected.
  6. On the General tab of the Restore Options dialog box:
    1. Select the name of the client computer from the Destination Client list.
    2. Select the Restore Control File check box.
    3. Click Advanced.

      The Advanced Restore Options dialog box appears.

  7. On the Restore tab, select the To Point-in-Time option and enter the date and time.

    Note: The selected time is specific to the CommServe computer's time zone.

  8. On the Recover tab, select the appropriate option.
    • If the database is not in NORACHIVELOG mode, select the To Point-in-Time option and enter the date and time.
    • If the database is in NORACHIVELOG mode and you want to perform a point-in-time backup from an incremental offline backup, select the Current Time option.
  9. When you restore the control file, on the Ctrl&SP Files tab:
    • Select the Restore From check box.
    • Select the Point-in-Time option and enter the date and time.

      You must restore the control files to a point-in-time later than or equal to the point-in-time set in the Restore tab.

  10. When the time zone is different for the client and CommServe computer, on the Options tab, in the Time Zone box, select the time zone of the client computer.
  11. To set the SnapProtect software to automatically move the database to the correct mode before the restore, select the Switch Database mode for Restore check box.
  12. Optional: Select the restore options.
    Expand All

    Set the Oracle Database ID

    The Oracle DBID is an internal, uniquely generated number that distinguishes the target database from the rest of the databases with the same name, in the recovery catalog. Oracle creates this number automatically when you create the database.

    You can use this option when:

    • There is no recovery catalog and you need to restore the control file or SP file from the autobackup
    • Multiple databases exist in the recovery catalog and you need to restore the control file

    On the Options tab, select the Set DBID check box.

    Open the database after the restore

    After a restore, set the SnapProtect software to automatically open the Oracle database. When the database is open, it records transactions.

    On the Options tab, select the Open DB check box.

    Reset the database and logs.

    By default, the database is automatically set to open, and the logs are reset.

    If you reset the logs to an open state, you can then reset the database.

    On the Options tab, select the Reset Database check box and one of the following Reset Log options:

    • To open the database without the RESETLOGS option, select None.
    • To open the database with the RESETLOGS option, select Yes.
    • To open the database with the NORESTLOGS option, select No.

    Perform a point-in-time restore of a database that was backed up in NOARCHIVELOG mode

    If the database was backed up in NOARCHIVELOG mode, enable the redo logs.

    When the no redo log is disabled, RMAN searches archived redo logs after applying the incremental backup data during a restore. When you set No Re-do Logs, RMAN restores the data from the incremental backup and not the archived redo logs.

    On the Options tab, select the No Re-do Logs check box.

    Prevent RMAN failovers to the previous backup for Oracle 10g databases or higher

    Select this option if you want to increase the speed.

    During restore operations, RMAN automatically looks for another copy of the backup file under the following circumstances:

    • A backup piece is corrupted or deleted
    • A MediaAgent is offline
    • A block in the backup is corrupted within the latest full backup

    If another copy is not available in the other copy, RMAN uses an older version of the file. When multiple channels are available for the same device type, RMAN automatically retries on another channel and searches all prior backups until it has exhausted all possibilities.

    On the Options tab, select the Disable Oracle Channel Restore Failover check box.

    Validate the restore

    Validation verifies that the backup copies of the data and logs are intact and usable for restores. RMAN simulates the restore job without the media to determine whether the restore can be successfully performed.

    When a validate job is complete, you can view the job log files to identify and correct any issues.

    On the Options tab, select the Validate check box.

    Configuring Pre- and Post-Processes

    Run scripts before or after the restore.

    Enter the full path for the pre-recovery and post-recovery script.

    You can choose to run the post-process script even if the restore job failed. You can use the post-process script to bring a database online or to release a snapshot.

    To pass the database instance name to the script, see Passing the Instance Name to Pre and Post Scripts for Databases.

    On the Pre/Post tab:

    1. In the Pre-Recovery Command box, type the full path name for the script.
    2. In the Post-Recovery Command box, type the full path name for the script.
    3. To run the post recovery process regardless of the job's outcome, select the Run Post Process for all attempts check box.
    4. In Windows configurations, specify the user that runs the process, in the Pre/Post Impersonation section, select one of the options:

      if the local account has permission to execute the processes on the destination client, select the Use Local Accounts option.

      To impersonate another user with permission, select the Impersonate User option and enter the credentials.

    Restore from a copy precedence

    If the backup is corrupted, restore from a storage policy copy instead of the backup.

    For more information on modifying the copy precedence of a storage policy copy, see Copy Precedence.

    On the Copy Precedence tab, select the Restore from Copy Precedence number check box and enter a copy precedence number.

    Restore from a specific backup

    On the Restore tab, select the By Tag check box and enter the tag.

  13. Click OK to close the Advanced Restore Options dialog box.
  14. Optional: View or customize the RMAN script that is generated from the selected options, see Viewing Oracle Restore RMAN Scripts and Customizing Oracle Restore RMAN Scripts.
  15. Click OK to close the Restore Options dialog box and start the restore.