PowerProtect Data Manager 19.14 Oracle RMAN User Guide

Performing centralized disaster recovery of an Oracle database or group

You can use the PowerProtect Data Manager UI to perform a centralized disaster recovery of an Oracle database including the spfile and control file. You can also perform a centralized disaster recovery of an Oracle Data Guard asset group.

1. In the PowerProtect Data Manager UI, select Restore > Assets and select the Oracle tab.

   The Restore window displays all the databases and Data Guard asset groups that are available for disaster recovery.

   You can click or on the top right of the window to switch between the group view and list view. The group view uses a tree view to show any Oracle Data Guard asset groups with their database assets. When you expand an asset group in the group view, you can see all the assets within the group. When you select an asset group, all the databases within the group are selected.

   In the list view, to filter the displayed list of assets if needed, you can click in the column heading Name, Status, Protection Policy, Host/Cluster/Group Name, Host Type, Data Guard Name, Data Guard Role, Application Name, Last Copy, Network, Protocol, or Backup Technology:

   • The Name and Application Name columns list the Oracle database asset names.
   • The Status column lists the status as Available, Deleted, or Not Detected.
   • The Protection Policy column lists the names of the protection policies for the assets.
   • The Host/Cluster/Group Name column lists the hostnames.
   • The Host Type column lists the host types as RAC or Standalone.
   • The Data Guard Name column lists the names of associated Data Guard groups.
   • The Data Guard Role column lists the Data Guard roles as Logical Standby, Physical Standby, Primary, Snapshot Standby, or Generic.
   • The Last Copy column lists the dates and times of the backup copies within the specified date and time range.
   • The Network column lists the networks that are available under the selected host or cluster.
   • The Protocol column lists the protocol as BOOSTFS or NFS when a protocol has been set.
   • The Backup Technology column lists the backup technology as Application Direct or Oracle Incremental Merge.

   NOTE:

   The listed assets are the assets that have at least one copy discovered by PowerProtect Data Manager. You can select only one asset for one Oracle host.

   Oracle software installed on the alternate host must be compatible with the source Oracle Server, as recommended by Oracle.

2. Select the check box next to the Oracle Server database asset or Data Guard asset group for restore, and click Restore.
   The restore wizard opens the Oracle Restore and Recovery window.
3. On the Scope page, select Disaster recovery, and then click Next.

   This option restores and recovers the selected database, including the spfile and control file, or the selected Data Guard asset group.

   You can use this option to restore the Oracle database to the original host or to an alternate host with a different database ID (DBID) than the original production host. The restore to the alternate host is usually used for test and development purposes.

4. On the Copy Selection page, select the backup copy to be restored from the list, and then click Next.

   To filter the displayed list of backup copies if needed, you can click in the column heading Created Time, Copy Type, Location, Backup Technology, Copy Status, Asset, Protection Strategy, or Storage Unit:

   • The Created Time column lists the dates and times when the backup copies were created.
   • The Copy Type column lists the backup copy types as Differential, Full, or Cumulative.
   • The Location column lists the backup copy locations as LOCAL or Local_Recalled.
   • The Backup Technology column lists the backup technology as Application Direct or Oracle Incremental Merge.
   • The Copy Status column lists the backup copy status as Available, Deleting, Deletion Failed, Deletion Failed (Agent Catalog), Cloud Tiering, Cloud Recalling, Replicating, Restoring, or Ready for Cloud Tiering.
   • The Asset column lists the database asset names.
   • The Protection Strategy column lists the protection strategy as All, Standby, Primary, or Standalone.
   • The Storage Unit column lists the storage units of the backup copies.
5. On the Location page, select the preferred type of restore, and then click Next:

   • Restore to original host—Specifies to restore to the original host with the displayed hostname.

     If the original host is part of a RAC cluster, select the available node hostname from the list.

   • Restore to alternate host—Specifies to restore to an alternate host.

     Select the alternate hostname from the list.

     NOTE:For an Application Direct restore, Oracle RMAN agent version 19.12 or later is required on the alternate host. For an Oracle Incremental Merge restore, Oracle RMAN agent version 19.13 or later is required on the alternate host.
6. On the Folder Location page, select one of the following options, and then click Next:

   • Restore to Root Level Folder

     Type the root level pathname in the Root Level Path text box.

     NOTE:

     For an Oracle Incremental Merge recovery, all the files related to the database are restored to this path, including the spfile.

   • Restore to original folder

     CAUTION:When you select to restore the Oracle data files to the original location, the original database is overwritten. Ensure that all the folder structure locations on the source host also exist on the target host.

     You can obtain a list of pathnames of the original directories by using any of the following methods:

     • Use the show parameter command from the original backed-up Oracle database.
     • Use the show parameter command from an Oracle instance using the restored spfile with the no mount clause or other suitable clause to start the temporary instance.
     • Use the create pfile command to generate the user-managed Oracle system parameter file.
     NOTE:Before you perform a disaster recovery, the directory pathnames listed by the show parameter command or used by the pfile must exist on the target host, owned by the Oracle user with proper permissions.

     For example, the pfile contains the following list of directory or file pathnames:

     *.audit_file_dest='/u01/app/oracle/admin/testsrc/adump'
     *.control_files='/u01/app/oracle/oradata/TESTSRC/controlfile/o1_mf_l3lybrr9_.ctl','/u01/app/oracle/fast_recovery_area/TESTSRC/controlfile/o1_mf_l3lybs7j_.ctl'
     *.db_create_file_dest='/u01/app/oracle/oradata'
     *.db_recovery_file_dest='/u01/app/oracle/fast_recovery_area'
     *.diagnostic_dest='/u01/app/oracle'

     In this example, before you perform the disaster recovery, the following directories listed in the pfile must exist or be created manually on the target host with proper permissions by the Oracle user:

     /u01/app/oracle/admin/testsrc/adump
     /u01/app/oracle/oradata/TESTSRC/controlfile
     /u01/app/oracle/fast_recovery_area/TESTSRC/controlfile
     /u01/app/oracle/oradata
     /u01/app/oracle/fast_recovery_area
     /u01/app/oracle

   • Restore to alternate folder

     To specify an alternate location, select Archive log, Control file, Datafiles, Fast recovery area, or Redo log from the menu, and then type the alternate location in the text box. For each additional alternate location, click +, select the file type from the menu, and type the alternate location in the text box.

     You must specify at least one alternate location if selected. For the control file, specify the full pathname, such as /u01/app/oracle/fast_recovery_area/sourcemay1/control01.ctl. You can specify a maximum of 31 archive log locations, five control file locations, one data file location, one fast recovery area location, and five redo log locations.

   NOTE:

   To ensure successful restores, the folders specified for the alternate folder locations and root-level folders must exist with the required Oracle user permissions.

   When you restore to an alternate host, the Oracle database audit folder must exist with the required permissions on the target host, similar to the source host. You must create the audit folder manually on the target host.

   When you restore to an alternate folder, all the required Oracle database directories must exist with the required permissions on the target host. You can obtain a list of pathnames of the original directories by using any of the following methods:

   • Use the show parameter command from the original instance.
   • Use the show parameter command from an Oracle instance using the restored spfile with the no mount clause or other suitable clause to start the temporary instance.
   • Use the create pfile command to generate the user-managed Oracle system parameter file.

   If you did not specify the alternate folder for one Oracle parameter, then the original directory must exist before you perform the disaster recovery. If you specified the new pathname for one Oracle parameter, then the new directory must exist before you perform the disaster recovery.

   For example, the original backed-up Oracle database uses "*.control_files='/u01/app/oracle/oradata/TESTSRC/controlfile/o1_mf_l3lybrr9_.ctl'". If you specify the alternate control file setting on the Folder Location page, then the directory that contains the control file must exist on the target host, created with proper permissions by the Oracle user. If you do not specify the alternate location for the control file, then the /u01/app/oracle/oradata/TESTSRC/controlfile directory must exist on the target host, created with proper permissions by the Oracle user.

   NOTE:The audit folder pathname in the audit_file_dest parameter setting can be changed on the SPFILE Options page.
7. On the Instance Details page, specify the required settings, and then click Next:

   • Oracle SID—Type the Oracle instance system ID (SID) in the text box.

     NOTE:The pathnames specified on the Folder Location page must match the Oracle directory naming method recommended by the Oracle.

     The SID must not be used by any existing Oracle instances on the host. The SID cannot be used by the files under the $ORACLE_HOME/dbs directory.

     If the recovery fails using the SID, you must manually shut down the restored Oracle database with the failure. You must also manually delete all files and directories created by the restore, including the files created under the $ORACLE_HOME/dbs directory that use the SID as part of the file name. Then you can use the SID to run the recovery again.

   • Oracle Home—Type the valid Oracle home pathname in the text box.

     NOTE:The Oracle home pathname must not include a final space or slash (/).

   • Credentials—Select one of the following options:

     • Use the credentials set at asset level or policy level for restore

       NOTE:Credentials at the asset level take precedence over credentials at the protection policy level.

     • Select existing credentials or create new credentials for restore

       Click Set Credentials, complete the fields in the Set Credential dialog, and then click Save.

   • Mount Protocol—Select BoostFS or NFS as appropriate. To perform a restore using the BoostFS protocol, the DD BoostFS package much be installed on the target client.

8. On the SPFILE Options page, specify the required settings, and then click Next:

   • SPFILE Options—Select this option to specify the restore of the spfile during the disaster recovery. This option is selected by default when the selected copy contains the spfile.

     If you do not select this option to restore the spfile for an Application Direct backup, you must create the spfile manually and start the database instance in no mount mode. If you do not select this option to restore the spfile for an Oracle Incremental Merge backup, the agent creates the pfile from memory.

     This option is disabled by default when the selected copy does not have an spfile backup. In this case, one of the following warning messages is displayed, based on the technology:

     Warning: The spfile is not found in the selected backup copy. Create the spfile or pfile manually and start the database instance in no mount mode.

     Warning: Oracle Incremental Merge will create the pfile from memory and restore will be performed using pfile.

     NOTE:

     Ensure that the spfile does not already exist on the target host. If needed, rename or delete the file on the target host; otherwise, the restore job fails.

     In some cases, the restore job from the UI might be reported as failed when pre-19.11 backup copies are selected. A known limitation exists in identifying specific backup pieces, such as an spfile and control file. The identification is required for a successful restore that is started from the UI.

     When the spfile option is cleared, only the Datafiles folder locations are used during the restore operation. Other added locations are ignored.

   • Configure spfile parameters (optional)—To change a default spfile parameter setting, type the parameter name in the Parameter text box and the parameter value in the Value text box. For each additional parameter setting that you want to change, click + and then type the parameter name and value in the new blank text boxes.

     For example, to specify the memory parameter settings sga_target=3000m and pga_aggregate_target=694m, type the following values in the text boxes:

     • Parameter: sga_target, Value: 3000m
     • Parameter: pga_aggregate_target, Value: 694m
     NOTE:

     If you specify the db_recovery_file_dest parameter setting, you must also specify the corresponding db_recovery_file_dest_size parameter setting. Ensure that the folder structure for the db_recovery_file_dest parameter is created with the required permissions on the target host. For example:

     • Parameter: db_recovery_file_dest, Value: /u02/app/sndusr/fast_recovery_area/sudb
     • Parameter: db_recovery_file_dest_size, Value: 8016m

     Ensure that the audit folder specified by the audit_file_dest parameter exists on the restore alternate host, owned by the Oracle user.

     If you do not specify the alternate pathname for the audit_file_dest parameter, the audit folder must be manually created with the same pathname used by the original backed-up Oracle database. The audit folder must be created with proper permissions by the Oracle user before you run the restore.

     If you specify the alternate pathname for the audit_file_dest parameter, that folder must be created with proper permissions by the Oracle user before you run the restore.

     For other parameters that use the folder pathname as the value, ensure that those folders are created with proper permissions by the Oracle user before you run the restore.

9. On the Database Recovery Options page, select the required options, and then click Next:

   • Restore To—Select Backup time (backup end time of selected backup copy) or Point in time.

     For Point in time, select one of the following options from the menu:

     • System Change Number—Type the System Change Number (SCN) in the text box.
     • Timestamp—Type the date and time in the text box, or click to display a calendar and select the date and time. Refer to the following guidelines on specifying the Timestamp option value.
     • Log Sequence—Type the log sequence in the text box.

     Use the following guidelines for the Timestamp option value that you specify for the restore:

     NOTE:It is recommended that you use the RMAN command list backup to determine the restore/recovery time point. The backup time displayed in the PowerProtect Data Manager copy information might not be in the recoverable time range.
     • If you have only one backup copy, you can determine the restore/recovery time range by the range from the data file backup completion time to the archive log backup next time. The following RMAN list backup command output shows the completion time of the data file backup and the archive log next time:

       BS Key  Type LV Size       Device Type Elapsed Time Completion Time     
       ------- ---- -- ---------- ----------- ------------ --------------------
       63      Incr 0  784.25M    SBT_TAPE    00:00:11     2023-APR-25 14:23:19
               BP Key: 63   Status: AVAILABLE  Compressed: NO  Tag: TAG20230425T142307
               Handle: ./PLCTLP-e2063cf4-d630-47a5-ad05-503780f2552b/Automated_data_TESTSRC_1826537176_1s1qg4uc_1   Media: S:ddve01.acme.com:/TESTSRC-ppdm01-77ecd/f091e90f
         List of Datafiles in backup set 63
         File LV Type Ckp SCN    Ckp Time             Abs Fuz SCN Sparse Name
         ---- -- ---- ---------- -------------------- ----------- ------ ----
         1    0  Incr 2683228    2023-APR-25 14:23:08              NO    /u01/app/oracle/oradata/TESTSRC/datafile/o1_mf_system_l3ly80cc_.dbf
       
       BS Key  Size       Device Type Elapsed Time Completion Time     
       ------- ---------- ----------- ------------ --------------------
       64      75.25M     SBT_TAPE    00:00:01     2023-APR-25 14:23:26
               BP Key: 64   Status: AVAILABLE  Compressed: NO  Tag: TAG20230425T142325
               Handle: ./PLCTLP-e2063cf4-d630-47a5-ad05-503780f2552b/Automated_arch_TESTSRC_1826537176_201qg4ut_1   Media: S:ddve01.acme.com:/TESTSRC-ppdm01-77ecd/f091e90f
       
         List of Archived Logs in backup set 64
         Thrd Seq     Low SCN    Low Time             Next SCN   Next Time
         ---- ------- ---------- -------------------- ---------- ---------
         1    18      2653863    2023-APR-20 20:00:38 2683276    2023-APR-25 14:23:23

     • If you have multiple backups and you want to restore/recover to the time before the latest backup, you can select any time after the first data file backup completion time and before the next time of the latest archive log backup.
     • If you want to restore the latest backup, ensure that you set the time to a point after the latest data file backup completion time and before the next time of the latest archive log backup.

   • Data Guard—Select the required option for a database restore in a Data Guard environment:

     NOTE:Ensure that the standby database is in sync with the primary database. You can restore a standby database backup to the primary database, or restore a primary database backup to the standby database.
     • None—By default, the None value is selected, which specifies to restore the database in a non-Data Guard configuration.
     • Primary—Select this option to restore the database as a primary database in a Data Guard configuration to a specified point in time.
     • Standby—Select this option to restore the database as a standby database in a Data Guard configuration to a specified point in time.

   • Database Options—Select any required options from the list:

     • Open database after recovery

       This option opens the database after the recovery. If the database is a container database (CDB), all the pluggable databases are opened.

     • Change DBID after recovery

       This option changes the DBID after the recovery. Select this option only when you also select to open the database after the recovery. When the database is unopened, the DBID cannot be changed.

       When you select this option for an Oracle Incremental Merge recovery, the database name is changed, based on the specified target SID.

       NOTE:

       When you select the Change DBID after recovery option, the asset is discovered automatically in PowerProtect Data Manager after a successful restore with the DBID change. When you do not select this option, the restored asset is not discovered automatically.

       Remove the SID entry from the EXCLUDE_SID parameter setting in the $RMAN_AGENT_HOME/config/rman_agent.cfg file.

     • Crosscheck backups

       This option specifies to crosscheck the backup before performing the disaster recovery. This option applies only when the application agent version 19.13 or later is installed on the target host. Selecting this option causes a delay in the display of progress updates for the restore job. Additional time is needed to crosscheck the backups from the Oracle RMAN catalog.

     • Dry run

       This option enables a job that creates the required RMAN restore scripts in the $RMAN_AGENT_HOME/tmp directory on the selected target host. The dry run procedure does not run a disaster recovery. You can use the scripts that the dry run creates to perform a self-service disaster recovery as required.

10. On the More Options page, specify the required options, and then click Next:
    • Set Stream Count—Type an integer stream count in the text box, if required. The default stream count is 4. The maximum stream count is 255.
    • Compressed Restore—To enable restore compression and reduce the impact on the network bandwidth, select Use PowerProtect DD Boost compressed restore.
    • Database Cleanup—To maintain the database in the mounted state when the recovery of Oracle Incremental Merge backups is partially successful, select Disable auto cleanup.
    • Troubleshooting mode—To enable troubleshooting logging, click the toggle button to change the setting to Enabled. Then the Log level option field appears.
    • Log level—If you enabled troubleshooting mode, select the preferred debug log level:
      • Info—Includes information such as status changes. This is the default log level for scheduled backups and restores.
      • Debug—Additional information that helps with problem diagnosis.
      • Trace—The most detailed amount of information for complex problem diagnosis.
11. On the Summary page:
    1. To ensure that the recovery settings are correct, review the information about the Scope, Location, Folder Location, Oracle Instance Details, SPFILE Restore Option, Database Recovery Options, and More Options.
       To change any settings, you can click Edit beside an information section. As an alternative, you can click Back to access a specific page and change the settings as needed.
    2. To start the centralized disaster recovery operation for the Oracle database, click Restore.
       A status message appears with a link to the Protection Jobs page where you can monitor the restore and recovery job.
12. Perform the appropriate actions after the restore job is completed:

    • If the restore job succeeded:

      • After a restore to an alternate host, it is recommended that you change the Oracle database name manually if the original backed-up database still exists.
      • Ensure that the Oracle redo log directory is updated to use the new SID with the pathname.
      • To restore the same backup to the same alternate host for multiple instances, you must rename the restored Oracle database. Then you can use the backup from the same Oracle database on this host with a different SID.
      • After a restore to the original host or an alternate host, restart the restored Oracle database

    • If the restore job failed, perform the following steps and then run the restore again:

      1. Manually shut down the restored Oracle database with the failure.
      2. Delete the restored files and files with the SID in the $ORACLE_HOME/dbs directory.
      3. Recreate the required directories as needed.
      4. On the More Options page in the PowerProtect Data Manager UI, enable the troubleshooting mode as described in the step 10 in this centralized disaster recovery procedure.