Re: What *really* happens at ALTER TBSP END BKP ?

"Tanel Poder" <tanel@@peldik.com> wrote in message news:3f05772a$1_1_at_news.estpak.ee...
> Hi!
>
> > take place naturally anyway. (It is true, however, that until the
headers
> > are synchronised, you are a little vulnerable to an unexpected instance
> > failure... Oracle would demand you perform a media recovery on
subsequent
> > startup to correct the SCN mismatch, when in fact no real media recovery
> is
> > required). Assuming you don't suffer from instance failures on a regular
> > basis, however, the extra checkpoint those books recommend is a bit
> > unnecessary.
>
> Actually you just have to issue end backup command to files in backup mode
> after instance crash.

Yes, I know that. But that won't help you in the situation I describe (otherwise I wouldn't have described it).

The issue is that you have already issued End Backup and *then* you have an instance crash. End Backup does NOT synchronise the datafile headers. So an instance crash in this situation means you have unsynchronised SCNs, and you do NOT have the bit in the control file to reset with 'alter database datafile xxxx end backup'.

At which point you simply have to initiate a recovery, although you can immediately cancel it.

Try and read what I actually wrote.

You're addressing an issue which I didn't. So now describe what you'd do when you've issued 'end backup', don't synchronise with a log switch or other form of checkpoint, and *then* you have a crash.

>Oracle just doesn't know whether the file is in backup
> mode because of a crash or because it's restored from older backup.

The point was, you're NOT in backup mode. Yet your headers are unsynchronised.

> No media recovery is needed if you don't lose your controlfiles
(checkpoint
> progress is also recorded in controlfile, thus Oracle knows it can recover
> using just online logs).
>
> Tanel.

~QM Received on Fri Jul 04 2003 - 08:11:23 CDT