Re: Why is undo backed with redo for GTT

From: Tanel Poder <tanel_at_@peldik.com>
Date: Sun, 2 Mar 2003 18:00:10 +0200
Message-ID: <3e622a5b_2@news.estpak.ee>

Hi!

> Tanel,
>
> that's exactly my understanding as well, but the question was more if
there is
> any cogent reason why redo is generated. If that could change in the
future,
> do we live with a less than optimal solution at the moment?
>
> Rene Nyffenegger

I think this is usual with any software that optmizing major new features take some time. Partitioning and global index maintenance improvements for example. There probably are better examples, which don't come into my mind for now.
I think it wouldn't be very easy or feasible to change rollback mechanisms that much, that they would actually start checking which rollback entry needs to be logged and which doesn't.

Another solution would be to have different type, non-logging rollback segments for GTT DML, but this would make overall architecture more complex. But, IMHO, we should keep everything as simple as possible (this applies to everything, not only Oracle or IS issues).

Cheers,
Tanel. Received on Sun Mar 02 2003 - 10:00:10 CST