Unitrends supports recovery of full and incremental backups. As with Oracle backups, the agent leverages RMAN to perform recovery operations. Oracle recoveries occur in two phases. First, the backup is extracted to the server’s storage and exposed as a CIFS share (/backups/rae/<client_name>/ <instance>). In the second phase, the client accesses the exposed CIFS share, and the RMAN is invoked to recover back to the Oracle database.
Consider the following before recovering Oracle data from the backup system:
• | Free space equivalent to twice the size of the backup is required on the Unitrends appliance for recovery processing. If adequate space is not available, the recovery fails. |
• | Recoveries are performed to the original database only. If you are recovering from a Unitrends appliance backup copy target and the original database is not available, you can recover from the target after performing a bare metal recovery of the Oracle client. See the Bare Metal Protection and Recovery Guide for details. |
• | Only one recovery or backup operation per database can run at any given time. |
• | For a given database, any backups initiated while a recovery is in progress fails. Once the recovery completes, backups can be run for the given database. |
• | For a given database, any recoveries initiated while a backup is in progress will fail. Once the backup completes, recoveries can be run for the given database. |
• | For item-level recovery, the backup is unfolded to a Unitrends appliance. From here you can recover items using an Oracle or third-party tool. After recovery, you must tear down the recovery object. See Recovering Oracle Backups for procedures and details. Subsequent backup or recovery operations for the given instance fail until the share has been manually torn down. |
• | For Oracle on Windows, recovery requires that the underlying file system has the same structure as when the database was initially backed up. For details, see KB 3354. |
The following recovery options are available:
• | Recover – Select a backup to recover all data in the backup group up to the point in time when the backup ran. |
• | Recover Files – Search a selected backup and choose specific files to recover. |
This procedure recovers a database to the original location. Note that the existing database is deleted from the Oracle instance as part of the recovery process.
1 | Select Recover and click the Backup Catalog tab. |
2 | Expand the Oracle asset and select the backup. |
3 | Click Recover. |
4 | Select the desired recovery options. |
5 | Click Save. |
To monitor the recovery, select Jobs > Active Jobs.
Notes:
• | In a successful recovery, status changes from Queued to Active to Successful. |
• | If the status displays as Canceled with a message Share is unavailable, the recovery cannot run because a share already exists for this instance. To determine what process is using the share and how to proceed, see Recovering Oracle Backups. |
Use this procedure to unfold the backup to a share on the backup appliance. Once unfolded, use an Oracle or third-party tool to recover desired items.
1 | Select Recover and click the Backup Catalog tab. |
2 | Expand the Oracle asset and select the backup. |
3 | Click Recover Files. |
4 | Click Confirm to create the recovery object and start the recovery. |
5 | Click OK to clear the Notice dialog. A row for this recovery displays on the File Level Recovery tab. |
6 | Select the Oracle recovery row and click Show Details to display the full path of the share: \\<ApplianceIP>\<InstanceID>. |
Note: Record the Network Path to access files to recover later.
7 | On the workstation to be used to recover files, map a network drive to \\<ApplianceIP>\<InstanceID>. |
8 | Recover the desired items using an Oracle or third-party tool. |
9 | Disconnect the network share once files are recovered. To disconnect the share, select the share and click Remove. |
10 | On the backup appliance, tear down the recovery object. |
Important! Tear down the objects soon as possible. Subsequent backups and recoveries cannot run for this instance until the object has been manually torn down.
Only one recovery or backup job per instance can run at any given time. If an Oracle backup or recovery operation fails with a share is unavailable message, a share already exists for this database. Review the following for additional information:
• | If a share exists for this database, it is in use by another backup or recovery process, or it has not been torn down after a previous recovery. |
• | If no active job is using the share, tear down the share. Disconnect any network drive mappings to the share before tearing it down. To tear down the share, select Recover, click the File Level Recovery tab, select the share, and click Remove. |
• | To view details on active jobs, select Jobs and click the Active Jobs tab. |
After files are recovered, the object remains until you tear it down. Because appliance resources are used to maintain the object, it is important to tear it down to ensure optimal performance.
To view Oracle recovery image
Select Recover and click the File Level Recovery tab.
To tear down the Oracle recovery object
1 | Disconnect any network drive mappings to the share before tearing it down. |
2 | Select Recover and click the File Level Recovery tab. |
3 | Select the recovery image to tear down. |
4 | Click Remove. |
Use this procedure if you are unable to recover from the backup system. Because Oracle recovery to an alternate server is not supported, this procedure requires that you perform a disaster recovery (DR) of the Oracle client to a new client that is directly attached to the backup copy target. See the Bare Metal Protection and Recovery Guide for details.
Once the client has been recovered, you perform an Oracle granular recovery. This operation is only supported on Oracle on Windows platforms running Oracle 10g, 11g or 12c.
Consider the following before recovering from a Unitrends appliance backup copy target:
• | Free space equivalent to twice the size of the backup is required on the local share for recovery processing. If adequate space is not available, the recovery fails. |
• | A replicated backup of the source client taken after the database was deployed is required. To perform DR using Windows integrated recovery, the backup copy target must have been run using version 7.3 or higher for BIOS-based clients, or version 7.4 or higher for UEFI-based clients. To perform DR using legacy Windows hot bare metal, the backup must be a bare metal backup. |
• | A backup copy of the database to recover is required. |
• | Only one recovery or backup operation per database can run at any given time. Any backup or recovery initiated while another is in progress fails. |
• | The Oracle backup is unfolded to a remote share. When you are finished recovering, you must tear down the share. Subsequent backup or recovery operations for the given instance will fail until the share has been torn down. |
1 | Perform a disaster recovery of the Oracle asset from the Unitrends appliance backup copy target. See the Bare Metal Protection and Recovery Guide for details. |
2 | Select Recover and click the File Level Recovery tab to determine if a share exists for this database. Review the following before proceeding: |
• | If no share exists for this database, proceed to the next step in this procedure. |
• | If a share exists for this database, it is in use by another backup or recovery process, or it has not been torn down after a previous recovery. You cannot perform the item recovery until the share is available. |
• | If no active job is using the share, tear down the share. Disconnect any network drive mappings to the share before tearing it down. To tear down the share, select it and click Remove. |
• | To view details on active jobs, select Jobs and click the Active Jobs tab. |
3 | Select Recover, and click the Backup Catalog tab. |
4 | In the Filter Backups window on the right, select Backup Copy. |
5 | Click Filter. |
6 | Expand the list of backups under the Oracle DB instance and select a backup to recover. |
7 | Click Recover Files. |
8 | Click the File Level Recovery tab. |
9 | Select the share and click Show Details. |
Record or copy the CIFS path as it displays in the File Level Recovery Details window.
10 | Log in to the target created in Step 1. |
11 | Open Windows Explorer. |
12 | In Windows Explorer, navigate to the CIFs path recorded in Step 9. |
13 | Open the file unitrends-<database>.env. Note the following values: |
• | Oracle SID |
• | Oracle Home |
• | Backup # |
14 | Open a command-line prompt. |
15 | Execute the following commands: |
# set ORACLE_SID=<SIDfromLastStep>
# set ORACLE_HOME=<HomefromLastStep>
# %ORACLE_HOME%\bin\rman target /
# shutdown immediate;
# startup nomount;
# restore controlfile from '<pathFromStep8>\unitrends-<database>.ctf';
# alter database mount;
# crosscheck backup;
# catalog start with '<pathFromStep8>\unitrends';
# list backup tag 'unitrends-<backup#FromLastStep>';
16 | Run the following commands: |
# restore database;
# recover database;
# alter database open resetlogs;
# quit;
The Oracle recovery is complete.
17 | On the Unitrends appliance backup copy target, disconnect any network mapping to the share and tear down the restore share. To tear down the share, select it and click Remove. |
Important! Tear down the share as soon as possible. Subsequent backups and restores cannot run for this instance until the share has been manually torn down.