Re: Corrupt rollback segment
Date: Tue, 7 Dec 2010 21:58:08 -0500
Thanks for all the helpful responses!
The story up to now:
We are performing restore/recover to a point in time where we know we were still good.
Oracle Support did come through for me, and worked under less than optimum support conditions (gov't agency doesn't allow logs or trace files to leave the environment)
The lesson learned: Even 20 Terabyte SANs can run out of disk space if you configure it to take a snapshot often enough!
Thanks for the help!
On Tue, Dec 7, 2010 at 9:30 PM, Kumar Madduri <ksmadduri_at_gmail.com> wrote:
> Hello Gus
> Will this help your scenario?
> at 3:59 PM, Gus Spier <gus.spier_at_gmail.com> wrote:
>> Solaris 10
>> Oracle 10.2.0.2
>> The database' SAN suffered a mild catastrophe very early this morning.
>> After the storage managers, sysadmins, and network types finally restored
>> service, we found that the database would not open
>> Review of the alert log highlights an ORA-600, ,,,, ..
>> Google search and MOS tell me that 4194 indicates that the redo block and
>> undo block do not match. In this case, the undo is at block 4 and the redo
>> is at block 7.
>> Further examination of alert log and trc files indicate my file# 7, block
>> 1777 is corrupt.
>> We have tried simple recovery, data file recovery, building a new UNDO
>> tablespace, but we still have an active rollback segment with a corrupt
>> A Severity One SR has been opened with Oracle Support, and been elevated.
>> We're still waiting for an update.
>> has anybody any useful suggestions?