Re: RMAN question

Brian Peasland <dba_at_remove_spam.peasland.com> wrote in message news:<3F93E15F.FE796EB_at_remove_spam.peasland.com>...
> > This is more of an opinion type question. Im reading the Backup and
> > Recovery book from Osborne. It states that its ok to have a recovery
> > catalog tablespace in an instance that has other data in it.
> >
> > This strikes me as a bad idea. I would think that you would want your
> > recovery catalog in a seperaet instance, then I would think you would
> > 'ideally' want a standby database that feeds off your RMAN catalog.
>
> Are you assuming that the Recovery Catalog is in the very database that
> you are backing up? If so, then that is indeed a bad idea. However, you
> can have a Recovery Catalog in a separate database that has other data,
> so long as the Recovery Catalog for that database is not in that
> database. For instance, I have two databases A and B that get backed up
> using RMAN. The RC for database A is in database B. And the RC for
> database B is in database A. Database A and B are used for other things
> as well. It should be noted that database A and database B are on
> separate physical servers.

Am I missing something? If both servers go down, is there enough info in the control file to restore one fer-shure? I'm a bit shy on this, since I've seen some pretty weird things without using a catalog, such as when you restore a few archived logs to a new location, the next time you backup the new location is included... so what will happen to a nocatalog restore after that? There's just so many possibilities, with concomitant surprises if you haven't tried any particular scenario. And of course, the situations that would bring down both servers are the ones where such weirdness will occur.

jg

--
@home.com is bogus.
http://www.quinion.com/words/weirdwords/ww-rhi1.htm