Tag Archives: corruption

RMAN recovery of lost or corrupted tablespace datafile

Oracle’s Recovery Manger (RMAN), originally released in Oracle 8, is a proprietary utility used for the backup and recovery of Oracle databases. Given a good Oracle RMAN backup, A DBA can recover from any disaster which may occur to their database, to include: loss of controlfiles, loss of datafiles, corruption of datafiles, loss of the entire database and even provides the ability to clone a database. In this article we give detailed instructions on the recovery of a lost or corrupted tablespace datafile in Oracle 9i, however these instructions will work in versions 10g and 11g.

1. Logon to the oracle server as the oracle software owner.

2. Set the ORACLE_SID and logon to sqlplus as sysdba.

[oracle@testdb02 scripts]$ su – oracle
Password:
[oracle@testdb02 ~]$ export ORACLE_SID=ORCL
[oracle@testdb02 ~]$ sqlplus ‘/ as sysdba’

SQL*Plus: Release 9.2.0.8.0 – Production on Thu Jun 12 20:03:06 2008

Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.

Connected to:
Oracle9i Enterprise Edition Release 9.2.0.8.0 – 64bit Production
With the Partitioning, OLAP and Oracle Data Mining options
JServer Release 9.2.0.8.0 – Production

SQL>

3. Query the dba_data_files table for tablespaces and datafiles. We are looking for a datafile which can be moved out of the location specified in the controlfile, resulting in tablespace corruption

SQL> select tablespace_name, file_name from dba_data_files order by tablespace_name;
TABLESPACE_NA FILE_NAME
————- ————————————————–
SYSTEM /u01/oradata/ORCL/system01.dbf
TOOLS /u05/oradata/ORCL/tools01.dbf
UNDOTBS1 /u04/oradata/ORCL/undo01.dbf
USERS /u05/oradata/ORCL/users01.dbf

SQL>

4. Shutdown the database with immediate.

SQL> shutdown immediate
Database closed.
Database dismounted.
ORACLE instance shut down.
SQL>

5. Exit SQL*PLUS and move to directory /u05/oradata/ORCL. Create a temp directory and move the datafile users01.dbf into the new temp directory.

[oracle@testdb02 ~]$ cd /u05/oradata/ORCL
[oracle@testdb02 ORCL]$ ls
temp02.dbf tools01.dbf users01.dbf
[oracle@testdb02 ORCL]$
[oracle@testdb02 ORCL]$ mkdir temp
[oracle@testdb02 ORCL]$
[oracle@testdb02 ORCL]$ mv users01.dbf temp/
[oracle@testdb02 ORCL]$ ls
temp temp02.dbf tools01.dbf
[oracle@testdb02 ORCL]$

6. Log back into sqlplus and startup your Oracle instance. Note: the error received referring to datafile users01.dbf, there exists a reference to this datafile in the controlfile, however Oracle could not locate the file. This results in the tablespace USERS being in an unusable inconsistent state.

[oracle@testdb02 ORCL]$ sqlplus ‘/ as sysdba’

SQL*Plus: Release 9.2.0.8.0 – Production on Thu Jun 12 20:48:14 2008

Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.

Connected to an idle instance.

SQL> startup
ORACLE instance started.

Total System Global Area 144670496 bytes
Fixed Size 741152 bytes
Variable Size 109051904 bytes
Database Buffers 33554432 bytes
Redo Buffers 1323008 bytes
Database mounted.
ORA-01157: cannot identify/lock data file 3 – see DBWR trace file
ORA-01110: data file 3: ‘/u05/oradata/ORCL/users01.dbf’

SQL>

7. To resolve the inconsistence, we must startup the RMAN utility and connect to both the target database and recovery catalog.

[oracle@testdb02 ORCL]$ rman

Recovery Manager: Release 9.2.0.8.0 – 64bit Production
Copyright (c) 1995, 2002, Oracle Corporation. All rights reserved.

RMAN> connect target sys/h1g0alz

connected to target database: ORCL (DBID=1373796530)

RMAN> connect catalog rman/rman@RMANDB

connected to recovery catalog database

RMAN>

8. From within the RMAN utility, issue the command to ‘restore tablespace ’. In this example the tablespace was USERS. The restore command replaces the tablespace’s datafiles in your database with the copy in your backup set.

RMAN> restore tablespace users;

Starting restore at 12-JUN-08

allocated channel: ORA_DISK_1
channel ORA_DISK_1: sid=15 devtype=DISK
channel ORA_DISK_1: starting datafile backupset restore
channel ORA_DISK_1: specifying datafile(s) to restore from backup set
restoring datafile 00003 to /u05/oradata/ORCL/users01.dbf
channel ORA_DISK_1: restored backup piece 1
piece handle=/u07/RMAN_BACKUP/ORCL/ORCL_657220129_1_26_1_0qjiooh1.dbf tag=BACKUP_FULL_ORCL params=NULL
channel ORA_DISK_1: restore complete
Finished restore at 12-JUN-08

RMAN>

9. From within the RMAN utility, issue the command ‘recover tablespace ’. The recover command applies all archive redo logs which have occurred since the last backup set.

RMAN> recover tablespace users;

Starting recover at 12-JUN-08
allocated channel: ORA_DISK_1
channel ORA_DISK_1: sid=14 devtype=DISK

starting media recovery

archive log thread 1 sequence 28 is already on disk as file /u01/oradata/ORCL/archive/ORCL_ARCH28.arc
archive log thread 1 sequence 29 is already on disk as file /u06/oradata/ORCL/archive/ORCL_ARCH29.arc
archive log filename=/u01/oradata/ORCL/archive/ORCL_ARCH28.arc thread=1 sequence=28
media recovery complete
Finished recover at 12-JUN-08

RMAN>

10. Exit the RMAN utility and enter SQL*PLUS, alter the database to an open status. You know that the tablespace USERS is now in a consistent state with the controlfiles when the database opens successfully

RMAN> exit

Recovery Manager complete.
[oracle@testdb02 ORCL]$ sqlplus ‘/ as sysdba’

SQL*Plus: Release 9.2.0.8.0 – Production on Thu Jun 12 21:03:11 2008

Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.

Connected to:
Oracle9i Enterprise Edition Release 9.2.0.8.0 – 64bit Production
With the Partitioning, OLAP and Oracle Data Mining options
JServer Release 9.2.0.8.0 – Production

SQL> alter database open;

Database altered.

SQL>

This completes successful recovery of a tablespace.

Larry J Catt, OCP 9i, 10g
oracle@allcompute.com
www.allcompute.com