|
Category :
Database
Resources -> Backup and Recovery
DB Version
:
Oracle 8 / 8i / 9i
OS Details :
Sun Solaris9
Recovering
from the loss or corruption of current and active redo log calls for
incomplete database recovery. Incomplete recovery is recovery to a
point in time to the past, before the media failure or corruption
occurred. Three types of incomplete database recovery exists in oracle
and they are change-based, time-based and cancel-based recovery.
Please be aware that you will lose all data changes made to the
database after the point to which the recovery was done. Also, the
database needs to be in archivelog mode for
the recovery to work.
Following steps need to be
done in order to do the incomplete recovery:
Step 1:
Startup the instance but do
not mount or open the database.
SQL> startup nomount
Step 2:
Query the database dictionary
for the current redo log number. The database recovery will be done
prior to this number:
SQL> select sequence# from
v$log where status = 'CURRENT';
Step 3:
Restore all the datafiles from
the most recent database backup. Do not restore the control files,
redo log files or init files from the backup.
Step 4:
Perform incomplete recovery.
Cancel-based recovery is probably the way to go even though it is a
manual process. When oracle tells you that it is ready to apply the
redo log sequence number that was obtained in step 2, cancel the
recovery.
SQL> recover database until
cancel;
Step 5:
Open the database and discard
the current redo log information.
SQL> alter database open
resetlogs;
Step 6:
If the redo log was corrupted
or lost due to a media failure and the database does not have the
minimum number of redo logs necessary for oracle to open the database
then drop the redo group whose files were lost and recreate it prior
to opening the database.
Step 7:
Take a full offline backup of the database and after the backup, make
the database available to the users. |
