After the failure of a storage system or a volume on a storage system, you can recover SQL Server databases that are mirrored using SnapMirror, as described in the following topics:
To be able to recover SQL Server databases using SnapMirror, you must complete the following disaster recovery preparation tasks:
You can configure SnapMirror to use a destination volume on the same storage system or a different, remote storage system. Whether the destination volume is on the same or a different storage system as the failed volume, the disaster recovery procedure is largely the same. For more information about configuring the Data ONTAP SnapMirror feature, see Data ONTAP Data Protection Guide for your particular version of Data ONTAP.
For each LUN on the SnapMirror source volume, note the relevant drive letter mapping on the SQL Server computer. You need to use the same mappings during the disaster recovery procedure when you use SnapDrive to connect to the corresponding LUNs on the SnapMirror destination volume.
The procedure entails recovering the relevant LUNs--the LUNs containing Snapshot copies of the databases to be restored--from the SnapMirror destination volume.
First use SnapDrive to connect to the corresponding LUNs in the SnapMirror destination volume. Then use the SQL Server management tool (Enterprise Manager or SQL Server Management Studio) to attach the database located on the SnapMirror destination volume.
If you cannot attach the database on the SnapMirror destination volume and the transaction log files are intact, follow the steps that describe how to minimize data loss by ensuring that SnapManager restore automatically back up the last active transaction log of the database.
If you succeed in recovering the necessary LUNs, you can use SnapManager to restore from the most recent backup set. Some of the details of this phase of the recovery procedure depend on the nature of the storage system or volume failure:
The recovery procedure includes special steps you must take if the transaction files were lost:
For descriptions of the two types of restore operations, see Types of SnapManager restore operations.
If you fail to disable this transaction log backup, subsequent SnapManager backup sets reside on a recovery path that is inconsistent with that of the database. Such backup sets cannot be applied to the database; attempts to restore the database from such backup sets results in failure, with the following error message in the restore log:
Failed with error code 0x800410df
If this occurs, perform the restore again, but do not apply transaction logs.
To recover SQL Server databases whose backup sets have been mirrored using SnapMirror, complete the following steps.
| 1 | ||
| ||
|
Go to Step 2. | ||
| 2 |
If any LUNs from the failed source volume still appear to be connected, disconnect them. |
|
| 3 |
Use SnapDrive to connect to the corresponding LUN or LUNs in the SnapMirror destination volume. Note ResultFor each mirrored volume, SnapDrive breaks the mirror and restores the LUN using the most recent Snapshot copy generated by SnapDrive or SnapManager. |
|
| 4 | ||
| 5 |
Use SQL Server Enterprise Manager or SQL Server Management Studio to attach the database located on the associated LUNs in the SnapMirror destination volume, as follows: |
|
|
Complete this procedure as described in Step 6. | ||
|
The database could not be accessed using SQL Server Enterprise Manager or SQL Server Management Studio |
Complete this procedure as described beginning with Step 10. | |
|
Complete this procedure as described beginning with Step 10. | ||
|
If you attached the database on the SnapMirror destination volume... |
||
| 6 |
The steps for restoring the database depend on whether the transaction log volume was lost: |
|
|
Complete this procedure as described in Step 7. | ||
|
Complete this procedure as described in Step 8. | ||
|
If you lost both the data files and the transaction log files of the database |
Complete this procedure as described in Step 8. | |
| 7 |
Run SnapManager and use the newest full database backup to perform either an up-to-the-minute restore or a point-in-time restore:
For more information, see the following topics:
|
|
|
Performing the restore if the transaction log volume was lost |
||
| 8 |
Disable the option to back up the transaction log prior to performing the restore: For more information, see SnapManager restore options. The reason you must disable this restore option is that the active transactions were lost due to the failure of the volume containing the transaction log. |
|
| 9 |
Run SnapManager and use the newest full database backup to perform a point-in-time restore. Note For a point-in-time restore, select the backup set, a combination of transaction log backups to be restored, or both. For more information, see the following topics:
|
|
|
If you did not attach the database on the SnapMirror destination volume... |
||
| 10 |
If you cannot attach the database on the SnapMirror destination volume and none of the transaction log files were lost, then--to reduce the loss of data--ensure that the last active transaction log of the database is automatically backed up by SnapManager Restore:
If any of the transaction log files were lost, no workaround is possible and you cannot minimize data loss. |
|
| 11 |
Use SnapManager Restore to automatically back up the last active transaction log of the database.
For general information about performing a restore operation, see Performing a restore operation. |
|
Copyright © 1994-2006, Network Appliance, Inc. All rights
reserved.
Part No. 215-01752_A0
File: dr7.htm
Last updated: 01-Sep-2006 07:41:24