Restoring Oracle Real Application Clusters database snapshots

Edit online

Follow the steps to restore an Oracle Restoring Oracle Real Application Clusters (RAC) database. In the example that is used in this task, an Oracle RAC database snapshot with two instances is restored.

Before you begin

Ensure that the following requirements are met:

The redo logs for the database are in a disk group that is not shared with any Oracle data files.
The Oracle RAC data files are stored on a supported storage system.

About this task

After the snapshot restore process completes, you can start the recovery of the restored RAC database. In the example that is used, the database is called MYDB, with two instances that are called MYDB1 and MYDB2. The procedure is for a complete recovery. For incomplete recoveries, refer to the Oracle documentation.

Procedure

Stop the database that is to be restored. Log on to the production server with the Oracle RAC instance owner user ID.

srvctl stop database -d MYDB
Restore a snapshot backup. Choose one of the following methods:
- Restore the last snapshot backup by specifying the command,
```
acsora -f restore 
```
- To restore an older snapshot backup, specify the backup ID of that backup as follows:
```
acsora -f restore -b backup ID
```
  where the backup ID is the ID of the snapshot backup that you want to restore.
To find an existing backup ID, type the following command: acsora -f inquire.
After the snapshot restore processing completes, you must recover the database.
The recovery type is specified in the profile file. For more information about recovery types, see DATABASE_CONTROL_FILE_RESTORE in ORACLE.
- If DATABASE_CONTROL_FILE_RESTORE YES is specified in the IBM® Storage Protect Snapshot profile file, an incomplete recovery is needed.
- If DATABASE_CONTROL_FILE_RESTORE NO is specified in the IBM Storage Protect Snapshot profile file, a complete recovery is needed.
Important: For all scenarios using snapshot technology with IBM Storage Protect Snapshot and using certain storage systems, the snapshot restore requires available space on the target storage pool so that it can restore the necessary volume. If there is not enough storage space available, you can increase the capacity on the requested storage pool, or free up some items that are using existing capacity. Check the message for the exact amount of storage space that is required.

For a complete recovery of the database, follow these steps:
1. Mount the database on the database instance where you want to run the recovery.
```
srvctl start instance -d MYDB -i MYDB001 -o mount
```
  If you want an incomplete recovery, do not proceed to the next sub step, see your Oracle documentation for information about incomplete recoveries.
2. Recover and open the database to complete the recovery operation.
```
SQL> recover database; SQL> alter database open;
```
3. Start all other database instances by entering the following command:
```
srvctl start instance -d MYDB  -i MYDB002
```
4. If you used normal or high redundancy, synchronize the failure groups that were not included in the restore operation.