If you were to lose controlfiles and a datafile, I don't think recreating
the controlfile would work. In this case, a recent backup of the
controlfile would allow you to recover to the point in time that the last
binary backup of the controlfile was done. So doing a binary backup of the
controlfile now and then would be a good safety measure -- especially when
the database structure changes (or it's back to the old structure).
Sybrand Bakker <postmaster_at_sybrandb.demon.nl> wrote in message
news:933185060.7063.0.pluto.d4ee154e_at_news.demon.nl...
> Hi Bill,
>
> Actually a control file is what it says: it has control information . The
> control information amongst others is the last SCN (system commit number)
> and the latest redo log sequence. It is used to determine what redo log
> files (archived or online) you need.
> If you wouldn't have used the using backup controlfile (which means: the
> controlfile is NOT current, it is from a backup) it would have started
> asking for redo log files.
> In this case because you used
> a) resetlogs, which did reset the redo log sequence number to 1 (or 0)
> b) using backup controlfile
> recovery didn't make any assumptions and simply read through all what was
> supplied.
> So, you did use the backup controlfile in a sense.
>
> Hth,
>
> Sybrand Bakker, Oracle DBA
>
> Bill Buchan <bill.buchan_at_ossian.net> wrote in message
> news:7nnbam$q4g$1_at_phys-ma.sol.co.uk...
> > I'm looking for help with the purpose of binary controlfile backups.
What
> > do you use them for?
> >
> > Specifically, can somebody explain what happened here?
> > I was testing a recovery situation where all I had left were the
datafiles
> > and the archive redo log files (ie. no controlfiles, on-line logs,
> password,
> > init.ora). The password files and init.ora were easily recreated.
> >
> > I then wrote a CREATE CONTROLFILE .. RESET LOGS script based on "backup
> > controlfile to trace" on a different database and editing it to match
the
> > datafiles names and directories that I had copied from tape.
> >
> > Once the instance was up and running again, I entered the command
> >
> > > ALTER DATABASE RECOVER UNTIL TIME '1999-07-28:10:30:00'; -- before
the
> > start of the on-line logs
> >
> > ORA-01610: recovery using the BACKUP CONTROLFILE option must be used
> >
> > > ALTER DATABASE RECOVER UNTIL TIME '1999-07-28:10:30:00' USING BACKUP
> > CONTROLFILE;
> >
> > ORA-00279: change 493146 generated at 07/28/99 05:00:04 needed for
thread
> 1
> > ORA-00289: suggestion: F:\ORADATA\ORCL\ARCHIVE\ORCLT0001S235.ARC
> > ORA-00280: change 493146 for thread 1 is in sequence #235
> >
> > > ALTER DATABASE RECOVER AUTOMATIC CONTINUE;
> > > ALTER DATABASE OPEN RESETLOGS;
> >
> > and the database opened...
> >
> > At no point did I have or apply a backup controlfile - what is the point
> of
> > "USING BACKUP CONTROLFILE"? And what is the point of taking a backup
> > controlfile if you don't seem to need it?
> >
> > Any explanations will be gratefully received!
> >
> > Many thanks,
> >
> > - Bill.
> >
> >
> >
> >
> >
>
>
Received on Mon Aug 02 1999 - 00:53:42 CDT
