Re: why r more archives generated during hot backup

From: Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>
Date: Mon, 23 Sep 2002 19:47:59 +0100
Message-ID: <amnnlt$7uf$1$8302bc10@news.demon.co.uk>

Connor,

You don't need to verify it, you've got the seminar notes (Session 03 Redo: slide 25 IIRC). You have to be careful about saying things like 'flushed out by dbwr' - it might give the wrong impression. If a block is written by DBWR it (usually) stays in memory, so it doesn't get a full copy image on the next changes. If it is flushed because __another_process__ wants to use the buffer and it is subsequently re-read then, as you point out, the first update thereafter writes two redo record again - the full block image, and then the normal redo vector.

--
Regards

Jonathan Lewis
http://www.jlcomp.demon.co.uk

Next Seminar dates:
(see http://www.jlcomp.demon.co.uk/seminar.html )

____England______September 24/26, November 12/14

____USA__________November 7/9 (MI), 19/21 (TX)

The Co-operative Oracle Users' FAQ
http://www.jlcomp.demon.co.uk/faq/ind_faq.html








Connor McDonald wrote in message <3D8F5CD0.3E69_at_yahoo.com>...


>

>I haven't physically sat down and verified, but I think its (only) the

>*first* change to the block since it is loaded into the cache.  Thus you

>might get (all in hot bkp mode)

>

>- change block   (whole block to redo)

>- change again   (just rows to redo)

>- flushed out by dbwr

>- change same block (whole block)

>- change again  (just rows)

>

Received on Mon Sep 23 2002 - 13:47:59 CDT