Re: Hot Backups

From: Graeme Sargent <graeme_at_pyramid.com>
Date: 1995/04/21
Message-ID: <3n706s$9a9_at_sword.eng.pyramid.com>#1/1


Judie Stroup (stroup_at_lmsc.lockheed.com) wrote:
: lmcgilvi_at_ionet.net (Laurel McGilvery) wrote:
: >We are currently backing up a 30 gig database to optical disk drives
: >using a combination of hot and cold backups. Has anyone out there run
: >hot backups backing up more than one tablespace at a time? I know the
: >book recommends against it.
: >Thanks.
: >
: >- Laurel
: >
: Laurel...
:
: We are currently performing hot backups and we do backup all the tablespaces
: at once (contrary to Oracle's recommendation). Of course, our current database is
: only about 2 gigabytes. We haven't experienced any problems with this method.
:
: judie

But if you look at the logic behind the recommendation you will see that it is faulty (or at least potentially faulty, depending on h/w config).

Oracle state that serial backup minimizes the time of exposure. This is not true providing your h/w config allows you to do parallel backup faster than you can do serial backup. (and if that is not the case, why would you want to do parallel backup anyway?).

Nevertheless, the correct procedure (IMHO) is a hybrid of the two procedures which Oracle document. In the parallel scenario you should issue the "END BACKUP" as soon as each tablespace completes, not batched after the last completion as the documentation implies.

What you should *not* do (and what I believe they really mean by their recommendation) is to do "parallel" backup of eight tablespaces to four tape drives (for instance). You should serialize two parallel backups of four tablespaces each, or better still parallelize four serial backups of two tablespaces each. (I oversimplify by ignoring the possibility of parallel datafiles within the same tablespace and implying tablespaces of equal size, but you should get the idea).

Admittedly, although parallel backup does not necessarily increase the *time* of exposure, it does increase the *scope* of exposure. But since the ability to backup in parallel faster than serially, implies the ability to restore in parallel faster than serially, this should not be an issue.

--
graeme
--
Disclaimer:	The author's opinions are his own, and not necessarily
		those of Pyramid Technology Ltd. or Pyramid Technology Inc.
---------------------------------------------------------------------------
      -m------- Graeme Sargent                 Voice: +44 (0)252 373035
    ---mmm----- Senior Database Consultant     Fax  : +44 (0)252 373135
  -----mmmmm--- Pyramid Technology Ltd.        Telex: Tell who???
-------mmmmmmm- Farnborough, Hants  GU14 7PL   Email: graeme_at_pyra.co.uk
---------------------------------------------------------------------------
    We have the technology.  The tricky bit is learning how to use it.
Received on Fri Apr 21 1995 - 00:00:00 CEST

Original text of this message