"Howard J. Rogers" <hjr_at_dizwell.com> wrote in message news:<41221dde$0$3928$afc38c87_at_news.optusnet.com.au>...
> Daniel Morgan wrote:
>
> > Howard J. Rogers wrote:
> >
> >> Daniel Morgan wrote:
> >>
> >>
> >>>bw wrote:
> >>>
> >>>
> >>>>I have inherited a 9.2.0 database where the only rollback segments
> >>>>that exist are what was created for SYSTEM. Can I get away with
> >>>>adding the needed rollback segments (RBS01... etc.) to the init.ora,
> >>>>creating them in the existing RBS tablespace? Once this has been
> >>>>completed do I then bounce the instance so they are recognized via the
> >>>>init.ora?
> >>>
> >>>SELECT segment_name
> >>>FROM dba_undo_extents;
> >>>
> >>>Unless this query returns no rows I'd suggest you get yourself
> >>>current with the 9i and 10g docs.
> >>>
> >>
> >>
> >>
> >> Now why would you get him to query that view and in that way?
> >>
> >> For meaningful output, you'd surely want him to select DISTINCT
> >> segment_name from dba_undo_extents.
> >>
> >> But then if you were going to do that, you might as well simply select
> >> segment_name from dba_rollback_segs... a view which is just as 'current'
> >> in 9i and 10g as any with the buzz-word 'undo' in it.
> >>
> >> Regards
> >> HJR
> >
> > Because it was the first query that came to mind that would quickly
> > indicate that there were undo segments.
>
> It was a rhetorical question. The point actually was: dba_rollback_segs
> tells you all you need to know about UNDO segments, too.
>
> In the rush to modernity that is 8i and 9i, let's not throw the 7 and 8.0
> babies out with the bathwater.
>
> > No other reason. And it pointed
> > the OP to the appropriate data dictionary view.
>
> My point exactly. It's a dictionary view, of course. And it has its uses,
> obviously. But it's not exactly "appropriate" for finding out which undo
> segments you have (if any) and what they're called.
>
> Would you query dba_extents to find out the names of the tables that exist
> in the database?
>
> That's a rhetorical question, too, btw.
>
> Regards
> HJR
Thanks to all for your help and education. as it turns out, the
UNDO_MANAGEMENT param is set to manual and we are running from the
spfile. Im still shaky on ROLLBACK_SEGMENTS as it relates to this
issue. UNDO_MANAGEMENT set to manual and ROLLBACK_SEGMENTS set to
null in the parameter file are confusing me. does oracle still
generate the needed rollback based on the calculation of the
TRANSACTIONS and TRANSACTIONS_PER_ROLLBACK_SEGMENT parameters?
Received on Tue Aug 17 2004 - 13:08:20 CDT
