PowerProtect Data Manager 19.14 Oracle RMAN User Guide

Restore failure due to FORMAT specification

An RMAN restore of the following types of backups might fail when the FORMAT specification contains a dot or period (.) character as a prefix:

• Self-service Oracle RMAN agent backup performed with the RMAN BACKUP command.
• Centralized Oracle RMAN agent backup through a PowerProtect Data Manager protection policy with a control file autobackup.

To prevent this issue, do not start the FORMAT specification with a dot or period (.) character.

Centralized restore issue with unchanged DBID

When you perform a centralized Oracle restore and select the option to change the database ID (DBID), the centralized restore might display the following error message:

Unable to complete changeDbid command: NID-00137: All datafiles that are not dropped should be readable.

In this case, the DBID is not changed after the restore. To change the DBID manually, use the instructions on the following Oracle site:

NID-00137 In Multitenant Database 12c (12.1.2) (Doc ID 2418171.1)

Centralized restore issue with audit file destination

When you perform a centralized Oracle restore, the restore job might fail with the following error message:

RMAN-04014: startup failed: ORA-09925: Unable to create audit trail file
Linux-x86_64 Error: 2: No such file or directory

To resolve this issue, complete either of the following actions:

• Create an Oracle database audit folder on the target host system, similar to the audit folder on the source host system.

• Specify the audit_file_dest parameter setting in the Configure SPFILE Parameters (optional) option in the PowerProtect Data Manager UI, as described in step 8 in Performing centralized disaster recovery of an Oracle database or group. For example, specify the Parameter and Value field settings in the UI as follows:

  • Parameter: audit_file_dest
  • Value: /u01/app/oracle/admin/source/adump

Disaster recovery might fail on same host with same SID

When you perform a disaster recovery to the same production host with the same SID by selecting an Oracle Incremental Merge backup copy, the disaster recovery might fail with the following type of error message:

Unable to perform live proxy restore as provided dbname '<database_name>' exists in the host.

As a workaround, complete the following steps:

1. Ensure that the database is in a deleted or shutdown state.
2. Ensure that the spfile and pfile of the database do not exist in the Oracle home directory.
3. Ensure that the database information is removed or commented out in the /etc/oratab file.
4. If you will perform the disaster recovery using the BoostFS protocol, check for a mounted NFS export. If an NFS export is mounted, unmount the NFS export manually.
5. Rerun the disaster recovery.

Instant access recovery might fail with NFS version 4

A centralized instant access recovery of an Oracle Incremental Merge backup that uses the NFS protocol version 4 might fail. The following error message appears on the Details tab in the Protection Jobs window:

ARO0507: PowerProtect Data Manager Oracle agent cannot perform the restore for the Oracle database.

As a workaround, run the following command from the DD console to enable the idmap service in DD:

nfs option set nfs4-idmap-active-directory enable

Instant access recovery issue might prevent cleanup through PowerProtect Data Manager UI

For an unknown reason, a centralized instant access recovery of an Oracle Incremental Merge backup might produce an incorrect report of an "unexpected" status for the recovery. This result prevents the automatic cleanup of the instant access recovery through the PowerProtect Data Manager UI.

As a workaround, you can complete the following steps to perform a manual cleanup of the instant access recovery.

1. To perform a query, run the ddutil command. For example, in the following ddutil command, fstdb is the source database name:

   ./ddutil -O OIM -n query -a SID=fstdb
   Live information found for database 'fstdb' with dbid:'2263439361':
   CLIENT : 'PLCTLP-bede9df3-3951-435e-bf86-bab32dea3090'.
   SID : 'door'
   DBID : '2263439361'
   BACKUP_PATH : '/IA-PLC-dcr-blrv003c081-cf147/PLCTLP-bede9df3-3951-435e-bf86-bab32dea3090/2263439361_87f1f629/full/restore/door_03062023151711'.
   BACKUP_HOST : '10.125.198.116'.
   STORAGE_UNIT : 'IA-PLC-dcr-blrv003c081-cf147'.
   NFS_MOUNT : 'TRUE'

2. Edit the rman_agent_oim.config file, using the output details from the preceding step. For example:

   [GENERAL]
   RMAN_AGENT_HOME = /home/oracle/opt/dpsapps/rmanagent
   BOOSTFS_INSTALL_DIR = /opt/emc/boostfs #only required in case of BoostFS
   CLIENT = PLCTLP-bede9df3-3951-435e-bf86-bab32dea3090
   [SID_door]
   BACKUP_HOST = 10.125.198.116
   STORAGE_UNIT = IA-PLC-dcr-blrv003c081-cf147
   ORACLE_HOME = /home/oracle/dual/oracle/product/12.2.0/dbhome_1
   NFS_MOUNT =TRUE #only required in case of NFS
   BACKUP_PATH =/IA-PLC-dcr-blrv003c081-cf147/PLCTLP-bede9df3-3951-435e-bf86-bab32dea3090/2263439361_87f1f629/full/restore/door_03062023151711
   OIM_TAG = 2263439361 #DBID output from ddutil query

3. To perform the manual cleanup, run the ddbmoim command from $RMAN_AGENT_HOME with the Oracle user. For example:

   ./bin/ddbmoim --dbname door -z config/rman_agent_oim.cfg -o livemount-cleanup -a ORACLE_OS_USER=oracle

   The latest PowerProtect Oracle RMAN Agent Administration Guide provides details about the ddbmoim command.

   NOTE:In case of NFS, if the unmount fails, manually unmount the path as the root user.

Oracle database becomes unavailable after centralized restore and PowerProtect agent restart or update

An Oracle database becomes unavailable for backups or other operations after the following operations occur:

1. You perform a centralized disaster recovery that restores an Oracle database.
2. At a later time, you restart or update the PowerProtect agent.

As a workaround, shut down and restart the Oracle database immediately after performing a centralized disaster recovery.

Troubleshooting issues with centralized and self-service restores of Data Guard backups

You might encounter the following issues with centralized and self-service restores of Data Guard backups:

• A centralized restore of a Data Guard backup that restores a stand-alone database might discover the newly restored database as a Data Guard database.

  As a workaround, when you perform the centralized restore in the PowerProtect Data Manager UI, add the following log_archive_config parameter setting on the SPFILE Options page under Configure spfile parameters (optional). The parameter value is two single quotes (''):

  Parameter: log_archive_config, Value: ''

• A self-service restore of a Data Guard backup that restores a stand-alone database might discover the newly restored database as a Data Guard database.

  As a workaround, after the restore is completed and before you register the database with PowerProtect Data Manager, perform the following steps:

  1. Run the following SQL command to unset the log_archive_config parameter. The parameter value is two single quotes (''):

     SQL> alter system set log_archive_config='';

  2. Run the following SQL command to stop the database:

     SQL> shutdown immediate;

  3. Run the following SQL command to start the database:

     SQL> startup;