Re: rman online backup with exclude undo tablespace

From: David Barbour <david.barbour1_at_gmail.com>
Date: Fri, 18 Jan 2008 17:17:45 -0500
Message-ID: <69eafc3f0801181417g575e69e6lc244f094403a2245@mail.gmail.com>


Check Metalink Note #*1013221.6. *

On 1/18/08, Prasad <p4cldba_at_gmail.com> wrote:
>
> As you mentioned I keep getting the same error no matter what i do .
> I tried already doing offline drop of undo datafile on mount while
> setting
> _*offline_rollback*_segments
>
> _*corrupted_rollback*_segments
>
> _*allow_resetlogs*_corruption
>
> even the oracle support suggest me to go ahead and copy a full backup . so
> I will take it as a learning and go back copy files.
>
> thanks everybody.
>
> -Prasad
>
>
>
>
> On Jan 18, 2008 1:00 PM, Jeremiah Wilton <jeremiah_at_ora-600.net> wrote:
>
> > Prasad,
> >
> > The error you are getting from open is that the UNDO tablespace
> > datafiles
> > are missing. Specifying _corrupted_rollback_segments will not solve
> > that
> > problem. If you would like to move on to the *next* error you are going
> > to
> > receive, then you can re-create the controlfile with the UNDO datafiles
> > removed, or you can OFFLINE DROP them from MOUNT mode.
> >
> > Notably, the method of saving time/space you devised is simply invalid.
> > You
> > need undo in order to create a consistent database. You can technically
> > open it without UNDO, but the result will range from inconsistent in the
> > best case, to constantly crashing in the worst.
> >
> > For those performing real recoveries of production data, I think this
> > group
> > is a great place to learn about extraordinary recovery methods, such as
> > how
> > to allow resetlogs/read only corruption, specify corrupted rollback
> > segments, and artificially advance the SCN. But for those like you who
> > are
> > willfully persisting in a flawed approach, the desire to help, at least
> > on
> > my part, is greatly reduced.
> >
> > Please just go back and get a copy of the UNDO tablespace which at 16G
> > is
> > really not that big. It doesn't matter if it takes 20 hours to make the
> > copy, as long as you can bring it all consistent with archivelogs. As
> > with
> > others, I recommend taking a more conventional approach to this project.
> > I
> > don't believe you will be providing your clients with a test system that
> > is
> > consistent with professional standards if you continue on your current
> > path.
> >
> > Regards,
> >
> > Jeremiah Wilton
> > ORA-600 Consulting
> >
> >
> > Prasad wrote:
> >
> > > Yes, it shows correctly printed in the alert log . but when I try to
> > do
> > show parameter it do not show . does
> > > that mean it is effective on the background? but if it is effective
> > then
> > when I try to open in resetlogs it
> > > should not complain about undotbs . but right now it complains.
> >
> >
> >
>

--
http://www.freelists.org/webpage/oracle-l
Received on Fri Jan 18 2008 - 16:17:45 CST

Original text of this message