Re: UNDO: 10g-style behaviour in

From: Michael Austin <>
Date: Fri, 03 Apr 2009 22:02:53 -0500
Message-ID: <s3ABl.4854$> wrote:
> Hi All,
> It is my understanding that UNDO works differently in 9.2 and 10.2.
> Size of undo datafiles is fixed, AUTOEXTEND is OFF.

> I wouldn't be surprised if Oracle "backported" 10.2-style behaviour in
> but it is not mentioned anywhere in the Release Notes for
> Patch Set.

I would be very surprised if they did... was the terminal/FINAL release of 9i and as of now, they are not back-porting any new fixes or features. It is a dead horse - time to move on to something that is still alive (for now...) 10gR2 (.05) is going to be the terminal release of 10g (whenever it hits the streets.) [This has been stated in several TARS filed at my previous employer] So, if you want to actually have a SUPPORTable database engine, I suggest starting down the 11g path. There are a LOT of new features that will help with true high availability - to the point that you may never need to take a complete downtime for your [RAC cluster] again. (you can do true "rolling" upgrades - even to completely new versions - to each node individually and still have the "database" accessible at all times.) Received on Fri Apr 03 2009 - 22:02:53 CDT

Original text of this message