Re: Recover index tablespace

From: Robert Klemme <shortcutter_at_googlemail.com>
Date: Tue, 20 Oct 2009 11:16:33 +0200
Message-ID: <7k5dflF38a025U1_at_mid.individual.net>


Please do not top post.

On 20.10.2009 07:08, Fabrice wrote:
> today, It happens nothing. It just a question for me.
> I had to restore my datafiles (index and data) there was some days. Problem
> with hardware. And now I'm looking for answers to somme questions.
> And what about corrupt or loss of the index tablespace (for example hardware
> failure or datafile corruption).
>
> And I down't know the best way to resolve this issue :
>
> - Can I just restore index datafile from my last cold backup and replay
> redolog archivied to get all my data and indexes consistents. (recover
> tablespace "tbs_index";)

If you have a proper DB backup basically it should not matter what tablespace is broken (apart from system maybe) - the backup will restore either properly.

> - Or do I have to recreate (by adding a new datafile) my index tablespace
> from nothing and launch a global rebuild of all the indexes. Can we do
> thant in production ?

I don't think so - unless your backup is fried, of course.

Btw, Richard Foote had a few interesting articles about the topic of separating data and index tablespaces on his blog. With reasonably sized databases thinking of index tablespaces as redundant information which can _easily_ be restored might buy you a ton of downtime (the rebuild likely is sufficiently slower than the restore).

Kind regards

        robert

-- 
remember.guy do |as, often| as.you_can - without end
http://blog.rubybestpractices.com/
Received on Tue Oct 20 2009 - 04:16:33 CDT

Original text of this message