Troubleshooting tips for IBM Storage Protect Snapshot for Oracle

Edit online

Resolving problems encountered when using IBM® Storage Protect Snapshot requires tasks specific to the native Oracle database environment.

If an error condition occurs during an IBM Storage Protect Snapshot event, there are several sources of information you can view to help determine what the problem might be. Be aware of the following information:

Make sure to increase the size of the following two Oracle options located in the $ORACLE_HOME/dbs/init(database_name).ora file:
```
sort_area_size = 10000000
sort_area_retained_size = 10000000
```
When using IBM Storage Protect Snapshot to back up an Oracle database, the target database being backed up cannot reside on the same volume group as the file system containing $ORACLE_HOME. Make sure that the Oracle Server does not share a volume group with the target database.

When performing a full offline backup of a database, the target database on the production server must be in startup mount state at the time acsora is issued. Otherwise it will not be possible to restore the resulting backup without performing recovery.

This RMAN script template will restore the database backed up offline as described in the previous paragraph. It restores control files, datafiles, and opens the database without any application of logs. This script must be started with the target database in a startup mount state:

run
{
allocate channel ch1 type 'SBT_TAPE' parms
'ENV=(TDPO_OPTFILE=<full path of tdpo.opt file>)';
set until scn = <Ckp SCN for backup being restored>;
restore control file to '<full path of 1st control file>';
restore control file to '<full path of 2nd control file>';
restore control file to '<full path of 3rd control file>';
alter database mount;
restore
(database);
sql 'alter database open RESETLOGS';
release channel ch1;
}

The database will in an open state and in a new incarnation after this script completes. All that remains is to issue the reset database command to RMAN and back up the database again since the previous backups are now rendered unusable since the database is in a new incarnation.

The

<Ckp SCN for backup
being restored>

value is the Checkpoint SCN listed for the backup being restored in the RMAN list backup command. For example, the Checkpoint SCN is 32024 in the following list:

List of Backup Sets
Key  Recid  Stamp  LV  Set Stamp  Set Count Completion Time
-------------------------------------------------------------
26081 4  469212393 0   469212319      5       06-AUG-02

List of Backup Pieces
Key  Pc# Cp#  Status  Completion Time  Piece Name
----------------------------------------------------
26082 1  1  AVAILABLE   06-AUG-02    05dvf74v_1_1

Lis of Datafiles Included
File  Name           LV Type Ckp SCN  Ckp Time
----------------------------------------------
1   /dev/rmyfilelv    0  Full 32024   06-AUG-02
2   /dev/rmyrollbklv  0  Full 32024   06-AUG-02
3   /dev/rmytemplv    0  Full 32024   06-AUG-02
4   /dev/rmyuserlv    0  Full 32024   06-AUG-02

Note that for an offline backup, the Checkpoint SCN should be the same for all of the datafiles.