Re: RAID-3 v. RAID-5

Ditto, RAID0 and RAID1 and RAID0+1.

And, was your 'non RAID' environment set-up correctly: say, striped disks, raw volumes, proper min and max pageahead, no write verify, correct stripe size, mullti bloclk read count, block size, etc., to make the best use of the drives you had? And did you test a number of variations of configuration, retuning both Oracle and the OS disk subsystem in between? No, thought not...

Oh, and have you assessed the impact of the controller with it's nice cache failing in the middle of a write?

MotoX

Chuck Hamilton wrote in message <3593c67f.12001044_at_news.dvol.com>...
>On Sun, 21 Jun 1998 05:07:33 GMT, shreterh_at_my-dejanews.com wrote:
>
>>Also, we've been hashing over whether RAID is slower per Oracle
performance
>>canon. After much reading and discussion, we concluded that the newer
cache
>>controllers eliminate mostly all the performance difference between RAID
and
>>nonRAID. In any event, we aren't showing disk problems right now, and we
can
>>always move redo, temp, rollbacks around later if we want to.
>
>Have you ever done any benchmarking between RAID-5 and RAID-3? If so,
>what were your results.
>--
>Chuck Hamilton
>chuckh_at_dvol.com
>
>If at first you don't succeed, sky diving isn't for you.
Received on Wed Jun 24 1998 - 04:51:16 CDT