Consistent Whole Database Backup Recovery

NOARCHIVELOG Database Recovery 

If a media failure damages datafiles in a NOARCHIVELOG database, then the only option for recovery is usually to restore a consistent whole database backup.

What is a whole database backup

Take a whole database backup of all files that constitute a database after the database is shut down to system-wide use in normal priority. A whole database backup taken while the database is open, after an instance crash or shutdown abort is inconsistent. In such cases, the backup is not a consistent whole database backup because the files are inconsistent with respect to a current point-in-time.

Whole database backups do not require the database to be operated in a specific archiving mode. A whole database backup can be taken if a database is operating in either ARCHIVELOG or NOARCHIVELOG mode. If the database is in NOARCHIVELOG mode, then the backup must be consistent (meaning the database is shut down cleanly before the backup).

To restore the most recent whole database backup to the default location:

1. If the database is open, then shut down the database.
• SHUTDOWN IMMEDIATE
2. If possible, correct the media problem so that the backup database files can be restored to their original locations.
• Restore the most recent whole database backup with operating system commands. 
• Restore all of the datafiles and control files of the whole database backup, not just the damaged files. 
• The following example restores a whole database backup to its default location:
• % cp /backup/*.dbf $ORACLE_HOME/oradata/trgt/ 
3. Because online redo logs are not backed up, you cannot restore them with the datafiles and control files. In order to allow the database to reset the online redo logs, you must first mimic incomplete recovery:
• RECOVER DATABASE UNTIL CANCEL;
4. Open the database in RESETLOGS mode:
• ALTER DATABASE OPEN RESETLOGS;