Re: Online backup: Backup online redologs?

"Charles Fisher" <Charles.Fisher_at_alcoa.com> wrote in message news:Pine.GSO.4.31.0105150914480.1016-100000_at_unknown...
> On Sun, 13 May 2001, Howard J. Rogers wrote:
 [Snip]
> However, with such a large userbase and the interaction of so many DBAs in
> the production of the script, your opinion on the merit of hot redolog
> backups is in question. Go and do some research.
>

It is not my opinion. It is in the doco, the training material, every Oracle book I've ever read, and there is a perfectly rational, well-researched and well-demonstrated explanation for why hot copies of redo logs are not to be taken.

> p.s. You are also wrong in saying "The absolute way of ensuring the
> current redo log is safe to copy is, of course, to shut the
> database down completely: now it's cold, and of course the copy
> results will be consistent." - If the database was brought down
> with a shutdown abort, the active online redolog will NOT be
> consistent.
>

Once again, you demonstrate your well-honed inability to understand a word I've been saying to you. The issue is what a pickle a dumb operating system and file system makes of copying a hot file. *That's* the inconsistency that I worry about, because it's that inconsistency that prevents the subsequent use of the copy. A redo log, copied with the database down, will be consistent, and I don't care HOW you shut the database down. The copy of that file can be used in a recovery, and all will be well.

And I demo as much every Backup and Recovery course.

End of discussion.

HJR Received on Tue May 15 2001 - 16:00:21 CDT