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: corrupt datafile

RE: corrupt datafile

From: Mark W. Farnham <mwf_at_rsiz.com>
Date: Tue, 15 Jun 2004 09:10:55 -0400
Message-ID: <KNEIIDHFLNJDHOOCFCDKGEDFEOAA.mwf@rsiz.com>


od (octal dump) and bp (binary patch) are generic Unix thingys which should have help files on your system.
-----Original Message-----
From: oracle-l-bounce_at_freelists.org
[mailto:oracle-l-bounce_at_freelists.org]On Behalf Of Terry Sutton Sent: Tuesday, June 15, 2004 2:04 AM
To: oracle-l_at_freelists.org
Subject: Re: corrupt datafile

Hi Mark,

We're trying some experimentation with underscore parameters. Not sure if it'll work out. I tried BBED, but all I could get was "BBED-00400: invalid blocktype (00)". I'd be interested in bp, but I don't know anything about it. And we're trying the other options. Except the $$$$$ one- the client isn't willing to spend money, which is how they got into this situation in the first place.

Stale data is not a concern. If we can get day old data, the client would be ecstatic.

Thanks,

--Terry

> if that outdated file is really all you've got, then you *might* be able
to:
>
> 1) cycle off your current undo and make a new one. (Unless you want it to
> try to roll back incomplete transactions. I'm not sure whether the
_corrupt
> undo and rollback inits are still operative.)
> 2) use od or that cool block dumper Steve Adams uses in his classes to
look
> at the headers
> 3) use bp to patch the headers to make the file look current (dump the
> headers of a few current "good" tablespaces to give you a clue, and/or
plead
> for mercy from someone who knows the block header details intimately
($$$$$
> usually can provide mercy).
> 4) turn on the event that keeps scanning after hitting corrupt blocks, and
> turn off block checking
> 5) give it a whirl and unload or export a table at a time.
>
> Of course your data will be stale and relational integrity will likely be
> screwed up unless there really was no action on the tablespace.
>
> So understand that unless that tablespace actually was cold and you get no
> block errors logged, then your data will not be perfect. Depending on
> whether you have parallel logical logs (non-Oracle application logs or
> something like that) from which you can replay the subsequent
transactions,
> you may be able to get most or all of what you need.



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 Tue Jun 15 2004 - 08:13:29 CDT

Original text of this message

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