Re: Recovery keeps asking for more logfiles

we had a situation like this once, and we discovered it was asking for one of the online redo logs. E-mail me, and I'll send you a link to what we documented from our experience.   Diana

Foggy wrote:

> "Richard Foote" <richard.foote_at_bigpond.com> wrote in message news:<Oc0g9.30427$g9.87989_at_newsfeeds.bigpond.com>...
> > Hi Foggy,
> >
> > Can be 100% sure but this sounds like the "David Attenborough Baby Bird
> > Syndrome".
> >
> > Love nature documentaries but always feel sorry for the parent birds trying
> > to feed baby chicks. After spending ages finding food, they return to the
> > nest, throw up in the baby chicks mouths only to have the chicks screaming
> > for more food. The bastards never shut up !!
> >
> > One of two scenarios here. Oracle behaves in a similar manner when you
> > recover a database and the control file is a backed control file or indeed
> > the control file is the "current" controlfile (but you only want to feed it
> > redo logs prior to the current time). Oracle uses the information in the
> > control file to determine when to end recovery in that all data files and
> > all control files must have consistent timestamps.
> >
> > In the case of the backed up control file, the control file is being
> > recovered in conjunction with the data files (as it's a control file from a
> > prior point of time) therefore Oracle has no way of knowing when to end
> > recovery (as the control file in this case doesn't contain the "current"
> > SCN). It just wants more redo logs and more redo logs and more redo logs ...
> >
> > In the case of it being the current control file, Oracle won't be satisfied
> > until it has applied all the redo logs and recovered all the data files
> > until such time that they reach the current SCN as documented in the control
> > file. That means it won't be happy until you have supplied the current redo
> > log which of course in your scenario won't yet have been archived. The
> > archive logs alone are not enough (the chicks haven't yet grown up enough
> > !!).
> >
> > I'm not sure if you wish to open this other database or not. If you do, you
> > must recover the database to the point where all the SCNs in the data files
> > *and* control files are all consistent (either by supplying all the redo
> > logs or by terminating an incomplete recovery).
> >
> > Sqwark Sqwark Sqwark Sqwark Sqwark Sqwark Sqwark ......
> >
> > Richard
>
> Hi Richard (and other people),
>
> thanks for your clear explanation.
> The datafiles are not in backup mode. I have tried recovery using a
> backup controlfile, and I have tried recreating the controlfile. The
> problem is not to open the database ( with recover database until
> cancel immediately followed by cancel), but this should be done
> automatically by a script.
> Replication can not be used here, because this customer wants the
> cheapest and most flexible solution. With replication, future database
> changes require additional work, and using the backup was considered
> to be a good solution.
> Does anyone have an idea to use a backup in this way?
>
> Regards, Johan
Received on Fri Sep 13 2002 - 14:04:49 CDT