Re: Question about redo size in dump file and view
From: <stevedhoward_at_gmail.com>
Date: Sun, 7 Jun 2009 14:50:48 -0700 (PDT)
Message-ID: <fcb1cd04-99ff-4af6-9611-e8bc34d7218d_at_g1g2000yqh.googlegroups.com>
On Jun 7, 3:10 pm, "Jonathan Lewis" <jonat..._at_jlcomp.demon.co.uk> wrote:
Date: Sun, 7 Jun 2009 14:50:48 -0700 (PDT)
Message-ID: <fcb1cd04-99ff-4af6-9611-e8bc34d7218d_at_g1g2000yqh.googlegroups.com>
On Jun 7, 3:10 pm, "Jonathan Lewis" <jonat..._at_jlcomp.demon.co.uk> wrote:
<snip>
> From the redo record you posted, the 488 is (roughly):
> Record header:
> 12 byte
> Change vector 1
> 140 bytes redo for table insert
> Change vector 2
> 132 bytes of undo header transaction table slot acquisition
> Change vector 3
> 80 bytes to clear the undo header transaction table slot
> Change vector 4
> 132 bytes for recording the undo vector for the table change
>
Jonathan,
Have you documented your findings/research methodology on this type of thing? If so, I would love to read it.
Thanks,
Steve Received on Sun Jun 07 2009 - 16:50:48 CDT