onmode -B
Hi,
onmode -B is a undocument command. Could I use this command to reduce thecheckpoint duration in IDS 9.30FC5.
It'll only 'sort of' work. Definitely a "It depends".
Prior to the introduction of non-blocking checkpoints, I did some
experimentation on just how much trying to flush the buffer pool prior =
to
issuing a checkpoint helped. At best it was a mixed results. The issu=
e is
that if the application is doing a lot of activity which updates stuff,=
then by the time that onmode -B is completed, the bufferpool is so dirt=
y
that the checkpoint still took a lot of time - and the net result was t=
hat
we were doing double flushing of a large majority of the pages.
Even if you are doing inserts and deletes, if you have a bunch of index=
es
on the table, then the index pages are being updated - so even that cre=
ates
a bunch of IO during the checkpoint.
=
=
From: "CHUAN LU" <luchuan@cn.ibm.com> =
=
=
=
To: ids@iiug.org =
=
=
=
Date: 04/07/2012 02:39 AM =
=
=
=
Subject: onmode -B [26683] =
=
=
=
Sent by: ids-bounces@iiug.org =
=
=
=
Hi,
onmode -B is a undocument command. Could I use this command to reduce t=he
checkpoint duration in IDS 9.30FC5.
***********************************************************************=
********
Forum Note: Use "Reply" to post a response in the discussion forum.
=