Re: Datafile corruption

I don't see the backup job can cause this problem. The worst is that the backup file is inconsistent and not useful for backup job.

If you are running in archiving mode, just using alter database to take the corrupt datafile offline and then physical delete it (Only do this if you are abolutely sure that it is corrupted and can not be repaired).

Then if all the log files since the datafile was created to the present are available,
and if your controlfile is current, then you can recover the datafile without a backup. In order to do this, you will have to use the CREATE DATAFILE clause of the ALTER DATABASE command: e.g. ALTER DATABASE CREATE DATAFILE 'oldfilename' AS 'newfilename'. This will create an empty file the same size as the missing file.
  Following this, you must perform media recovery on the empty file. During media recovery of the empty file, all archived and online redo logs must be available to be applied to the file so that it can be brought back to a transaction consistent state

>List Gurus, Today when I came to work I observed that one of my datafiles
was
>corrupted. I am trying to figure out the reasons why this happened. I have
>attached a part of the alet log at the bottom of this message. Any help
with
>this greatly appreciated. Few insights into this is that, very recently we
>started a new backup process. The way it is done is a master batch file
calls
>oracle stop batch file, then the backup batch file that copies the data
over
>to backup drive and then the oracle start batch file. The oracle start and
>stop are done using NT services "net start". Could these processes be
running
>over each other .... Any help, insights into this issue are greatly
>appreciated.
>
>Sanjay.
>
>Alert Log:
>
>SNP1 started
>Mon Nov 23 20:31:21 1998
>alter database "projr" mount exclusive
>Mon Nov 23 20:31:21 1998
>Successful mount of redo thread 1.
>Mon Nov 23 20:31:21 1998
>Completed: alter database "projr" mount exclusive
>Mon Nov 23 20:31:21 1998
>alter database "projr" open
>Mon Nov 23 20:31:24 1998
>Errors in file D:\ORANT\rdbms80\trace\proLGWR.TRC:
>ORA-01110: data file 6: 'H:\ORANT\DATABASE\PRO\PROINDX1.ORA'
>ORA-01114: IO error writing block to file 6 (block # 1)
>ORA-27072: skgfdisp: I/O error
>OSD-04008: WriteFile() failure, unable to write to file
>O/S-Error: (OS 5) Access is denied.
>
>Mon Nov 23 20:31:24 1998
>Thread 1 opened at log sequence 45258
> Current log# 1 seq# 45258 mem# 0: H:\ORANT\DATABASE\PRO\LOG1.DBF
> Current log# 1 seq# 45258 mem# 1: H:\ORANT\DATABASE\PRO\LOG1B.DBF
> Current log# 1 seq# 45258 mem# 2: H:\ORANT\DATABASE\PRO\LOG1C.DBF
>Successful open of redo thread 1.
>Mon Nov 23 20:31:24 1998
>SMON: enabling cache recovery
>SMON: enabling tx recovery
>Mon Nov 23 20:31:27 1998
>Completed: alter database "projr" open
>Mon Nov 23 20:33:10 1998
>Thread 1 advanced to log sequence 45259
> Current log# 2 seq# 45259 mem# 0: G:\ORANT\DATABASE\PRO\LOG2.DBF
> Current log# 2 seq# 45259 mem# 1: G:\ORANT\DATABASE\PRO\LOG2B.DBF
> Current log# 2 seq# 45259 mem# 2: G:\ORANT\DATABASE\PRO\LOG2C.DBF
>Mon Nov 23 20:33:14 1998
>Errors in file D:\ORANT\rdbms80\trace\proCKPT.TRC:
>ORA-01171: datafile 6 going offline due to error advancing checkpoint
>ORA-01122: database file 6 failed verification check
>ORA-01110: data file 6: 'H:\ORANT\DATABASE\PRO\PROINDX1.ORA'
>ORA-01208: data file is an old version - not accessing current version
>
>-----------== Posted via Deja News, The Discussion Network ==----------
>http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
Received on Tue Nov 24 1998 - 19:18:18 CST