Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: why did this recover scenario fail ?

Re: why did this recover scenario fail ?

From: DNP <High.Flight_at_btinternet.com>
Date: 2000/03/30
Message-ID: <38E3D5CF.F02@btinternet.com>#1/1

Editing the control file to leave out the lost datafiles may sound O.K. on the surface, but there will still be a data dictionary / physical storage settings conflict as soon as you try and mount. The data dictionary still knows about the tablespaces 'owning' the affected datafiles, and so you must tell Oracle to lose its association with the datafiles and then the tablespaces using its SQL commands. In this way the control file is properly 'informed' and should not cause spurious media recovery suggestions on mounting or opening.

David P.

Oracle Certified DBA.


Jason Kratz wrote:
>
> Not sure about the system tablespace but mounting the database and doing a
> backup controlfile to trace would only give you a text version of what your
> system currently had (ie: the temp and index stuff included). unless it was
> edited to not include those things you'd just be recreating a control file
> for the same setup.
>
> Jason
>
> "NeilC" <nospam.neilc-olos_at_btinternet.com> wrote in message
> news:38E372FC.C6C7CDE5_at_btinternet.com...
> > Has anyone an explanation of what went wrong with this recovery scenario
> > ?
> >
> > Oracle 8.0.5 on NT
> > noarchivelog mode
> >
> > A small test database that had just been created with 5 tablespaces
> > (system, temp, rbs, data and indx) was populated with data - all ok.
> > Shortly afterwards one of the 2 disks failed. This disk contained the
> > temp and indx tablespace and the second members of a set of 4 redo log
> > groups and the second of two controlfiles.
> > So, I was left with one control file, system, rbs and data tablespaces
> > and 4 redo logs.
> > As the temp tablespace wasn't needed, and the indx tablespace wasn't
> > needed (indexes could be re-created) I thought I might be able to
> > re-create the controlfile to only include the 4 redo log members, the
> > three tablespaces datafiles and the current control file.
> > So, I mounted the database issued a backup controlfile to trace and
> > recreated the control file - all seemed to be going well.
> > I now thought i'd be able to open the database, reset the redo logs and
> > then go about recreating the temp and indx tablespaces. But, when trying
> > to open the database, it wanted to do media recovery on the system
> > tablespace. I tried an recover database until cancel, but it said that
> > the datafile would still require recovery.
> > So, disappointedly, I gave up, re-created the DB and reloaded my data.
> > What did I do wrong ?
> >
> > Cheers
> >
> > Neil
Received on Thu Mar 30 2000 - 00:00:00 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US