Re: Oracle 7 Crash on NT4 !

Hi Eric,

In the middle of your error messages, when you are backing up the database, you get this:

ORA-1128 signalled during: alter tablespace KCA_RBS01 begin backup...
>Tue Mar 16 20:33:55 1999
>alter tablespace KCA_RBS01 end backup
>ORA-1142 signalled during: alter tablespace KCA_RBS01 end backup...
>Tue Mar 16 20:33:55 1999

At the specified time of failure, you get this:

Automatic datafile offline due to write error on file 8: D:\ORANT\DATABASE\CONTAUTH\ROLLBACK\ROLLBACK02.DBF KCF: write/open error dba=0x80008e4 block=0x8e4 online=0

     file=8 D:\ORANT\DATABASE\CONTAUTH\ROLLBACK\ROLLBACK02.DBF
     error=9205 txt: 'OSD-04016: Error queuing an asynchronous I/O request.

Which is OracleSpeak for 'Damned if I can write to that file! I am going to take it offline so that you can fix it.'. Since this datafile is in a rollback segment tablespace, this error happens when someone runs SQL against the database (or an export, perhaps?) and the database cannot write to the rollback segments.

Later, when you try to backup the database, you get this:

ORA-1128: Oracle's way of telling you that a datafile in that tablespace is offline, and therefore the tablespace cannot be backed up.

ORA-1142: Oracle's way of saying that this tablespace cannot be taken out of hot backup since it is not in hot backup.

Looks like you have a corrupt datafile, in your rollback segment tablespace.

Create some temporary rollback segments in another tablespace, drop the rollback segment tablespace with the INCLUDING CONTENTS option (make sure that it is only a rollback segment tablespace, please! No data in there), delete all the datafiles in this tablespace from the disk, recreate the rollback segment tablespace, recreate the rollback segments, drop the temp rollback segments, and viola! problem is fixed, we hope.

Cheers,

Jack Silvey
Oracle Certified Professional

If
Merlin Team wrote in message <7cr7t4$d2m$1_at_news1.cableinet.co.uk>...
>Hello there.
>
>I hope someone can point me in the right direction with a problem on Oracle
>7.3.2.3.1. on NT4.
>
>It would appear that something goes wrong at approx 23:09 every Tuesday !
>There is a scheduled hot backup done at 20:30 every night which lasts about
>1 hour at most but
>the database appears to have started up ok after that.It then dies and
 shuts
>down the instance.
>The other 3 databases on the same server continue ok. The only way I have
>found to recover the
>situation is to restore from the Tues backup.
>
>I have included parts of various log/trace files below.
>
>TIA
>
>Eric Penninton
>eric_merlin_at_hotmail.com

<snip> Received on Fri Mar 19 1999 - 00:00:00 CET