Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Mailing Lists -> Oracle-L -> Re: BAARF - true life story

Re: BAARF - true life story

From: Alexandre Gorbatchev <agorbatchev_at_amadeus.net>
Date: Wed, 11 Aug 2004 23:29:41 +0200
Message-ID: <OFD3524F40.5E9F413A-ONC1256EED.00758145-C1256EED.007613B4@amadeus.net>


As few ppl. mentioned here the fastest way would be a physical standby database with Data Guard or without (aka 8i style). I can only confirm that this approach was used at our site in production to move a critical database to another node and (most important) upgrade it to 9.2.0.5 at the same time in few minutes (this is another story, but Oracle is now considering this approach as official way of upgrade for HA environments). We didn't use Data Guard since this is very sensitive procedure and better be handled manually (well, with our own scripts). Regards,
Alex

From: <solbeach_at_cox.net>@freelists.org on 07-08-2004 15:18 AST Please respond to oracle-l_at_freelists.org Sent by: oracle-l-bounce_at_freelists.org

To:
<oracle-l_at_freelists.org>

cc:

Subject:
BAARF - true life story

I just joined this site.
Less than 6 weeks ago they had a problem with one of their Oracle DBs. It seems it was running on a RAID-5 volume. A disk died. The hot spare got automagically included, but nobody was notified or noticed this happening. Then 2nd disk failed. It was some time after this that errors started and one of the *DBF had gotten corrupted. They have been able to reload the records which had been sitting on the bad block. So we think that from a logical (data) standpoint the problem has been resolved. The DBV utility no longer reports any bad blocks.

We now have to move this whole database (178GB) to a new system, with minimal downtime.

We have worries, doubts & concerns about just doing an OS copy from the old to the new system. Can there still be 1 or more bad blocks in the tablespace that just are not being used now? If the OS (Linux) has tagged & remapped bad blocks, will an OS copy allow them to be "left behind"? We doubt we have a big enough downtime window to allow for either an expotr/import to RMAN duplicate operation.

What would YOU do given this history & requirements?



Please see the official ORACLE-L FAQ: http://www.orafaq.com

To unsubscribe send email to: oracle-l-request_at_freelists.org put 'unsubscribe' in the subject line.
--

Archives are at http://www.freelists.org/archives/oracle-l/ FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html


Please see the official ORACLE-L FAQ: http://www.orafaq.com

To unsubscribe send email to: oracle-l-request_at_freelists.org put 'unsubscribe' in the subject line.
--

Archives are at http://www.freelists.org/archives/oracle-l/ FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html
Received on Wed Aug 11 2004 - 16:25:46 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US