Re: Disk MTBF

I have found that using mirrored disks to be a good solution for short term storage for the Oracle backup's. I usually use disks for my nightly backups because they are the fastest way to get backups completed. After the Oracle backups are completed our system administrator backs up the Oracle files plus other Unix files to tape. With this scenario one is protected ether way and is able to get back on-line much quicker. Also, if one has to recover, I don't have to want around for the tape to pull it back to disk ... a very slow process!

Hope this helps.

-rws

"Fred A G" <nospam_at_allowed.localhost> wrote in message news:5hBn6.5593$hi2.16367_at_nntpserver.swip.net...
>
> "Roger Menhennett" <r.menhennett_at_lmco.com> wrote in message
> news:3A9BE9CD.B783FE73_at_lmco.com...
> > Am wondering what the probability of disk failure is in the following
> > scenarios for the first year, second year, etc. (assume one machine
 with
> > a
> > disk mtbf of 50000hr.).
> >
> > 1. One disk.
> > 2. Six disks.
> > 3. Six disks configured as raid1.
>
>
> > Am concerned with how much quantifiable confidence one can put in a
> > raid device to substitute it for backups. Besides device failure,
>
> None, nada, nil! A raid device is not a substitute for backup (where you
> make a good copy of important data to tapes and place the tapes in the
> media safe(s))!
>
> Mirroring/redundacy in raid is for "protection" against disk failure,
> nothing more. A six disk simple mirroring raid-1 (3 <-> 3) can stand
> loosing one to three disks. Make sure you spread disks across several
> HBAs too, just in case...
>
> Regards
> /Fad
>
>
Received on Fri Mar 02 2001 - 10:57:29 CST