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

Home -> Community -> Usenet -> c.d.o.server -> Re: Question about 'split blocks'

Re: Question about 'split blocks'

From: pady srinivasan <pady_at_imncorp.com>
Date: 1997/08/10
Message-ID: <33EE1156.1E45F246@imncorp.com>#1/1

This is a multi-part message in MIME format.

--------------4C4CB614871F6ED348452475
Content-Type: text/plain; charset=iso-8859-1
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit

pady srinivasan wrote:

> I couldnt understand this concept in the internal operation of hot
> backups. In the 'Oracle Backup and Recovery' book by Rama Velpuri,
> this
> is the explanation given for Oracle logging entire block images during
>
> hot backups...
>
> " If the Oracle block size is a multiple of OS blocks, then depending
> on
> how the OS copies blocks, it is possible for a hot backup to contain
> an
> inconsistent version of a given data block. For example, if a block is
>
> updated on disk between the reads, the copy in the backup file could
> be
> useless since the front and back halves of a block may be written at
> different times."
>
> Question: I can understand that the OS can copy OS blocks at different
>
> times and in the meantime, the Oracle block could have been changed.
> But
> since this change is in the redo, why should the hot backup ( being an
>
> OS copy ) contain an inconsistent version of the data block OR in what
>
> way will this cause a problem during recovery ?
>
> Thank you. If possible, please send me a reply to pady_at_imncorp.com
>
> Pady Srinivasan
> Integrated Medical Networks
> pady_at_imncorp.com
>
> --
> MZ
>
>
> -------------------------------------------------------------------------------------------------------------
>
> pady Srinivasan <pady_at_imncorp.com>
> Oracle DBA/UNIX sysadmin
> Integrated Medical Networks
>
> pady Srinivasan
> Oracle DBA/UNIX sysadmin <pady_at_imncorp.com>
> Integrated Medical Networks
> Netscape Conference Address
> Netscape Conference DLS Server
> Additional Information:
> Last Name Srinivasan
> First Name pady

I think I found the answer to this in the same book. The answer is

" To verify the consistency of the block before recovery, the version number at the beginning of the block is compared to the version number at the end of the block to determine whether the block has been split during a hot backup. If the version numbers are not the same, a consistent version of the block is required and the before image of the block in the redo is copied to disk before applying redo changes."

Thanks

Pady Srinivasan

--
MZ


--------------4C4CB614871F6ED348452475
Content-Type: text/x-vcard; charset=us-ascii; name="vcard.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for pady  Srinivasan
Content-Disposition: attachment; filename="vcard.vcf"

begin:          vcard
fn:             pady  Srinivasan
n:              Srinivasan;pady 
org:            Integrated Medical Networks
email;internet: pady_at_imncorp.com
title:          Oracle DBA/UNIX sysadmin
x-mozilla-cpt:  ;0
x-mozilla-html: FALSE
end:            vcard


--------------4C4CB614871F6ED348452475--
Received on Sun Aug 10 1997 - 00:00:00 CDT

Original text of this message

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