Re: Raid5 Vs Raid0+1 -- Raw Vs Solaris 9 Concurrent Direct IO UFS

From: Paul Drake <discgolfdba_at_yahoo.com>
Date: Thu, 26 Aug 2004 00:22:51 -0700 (PDT)
Message-ID: <20040826072251.61984.qmail@web20424.mail.yahoo.com>

Deen Dayal <deen.dayal_at_dol.state.nj.us> wrote:

> We are planning to move our current database to new
> Hardware Sun 12k +
> Hitachi SE 9970.
>
> Our current database runs on Sun A5000 array with
> Raid0+1 on raw devices.
> Sun engineer who is here at our site for
> implementation of Hitachi SE9970
> suggests that we should go Raid5 ( 3 + 1P ) in the
> parity group as RAID5
> gives better read performance than raid0+1 and any
> writes are going to the
> cache any way, so we should not be worried about
> write performance. There is
> 16GB cache on the Hitachi.

That salesperson made their quota right there. Their golf pro should be sending you a fruit/cheese basket.

> Database is going to be striped across 10 parity
> groups and each parity group consists of 4 physical
> disks.

no wonder my !@#$%^ nj taxes are so damned high.

4 drive raid 5 volume?
ask him to do the math.
ask him what you optimal db_file_multiblock_read_count should be.

if he would have said 5 drive RAID 5 volume, I would have at least listened further.

does he have the slightest clue as to _your_ application's block size (e.g. db_block_size).

the SA was assimilated.
hopefully, the vendor took him out on a most enjoyable venture to a city such as Montreal.
One full of culture.

beware of a small, slimy creature with large teeth tearing open his abdominal wall (from the inside) to escape to gestate into an 8 foot tall, mean, human chewing alien.

As with database re-orgs, new hard drives that are say Ultra 320 SCSI (320 MB/sec interface), replacing existing U2W SCSI (80 MB/sec interface) - where the new, sexier models outperform the older, more reliable but closer to MTTF, 4 year old models, masks the true issue.

At least you aren't comparing apples.

baarf.

check out the paper at hotsos.com on "aligning database block and stripe size boundaries".

Please see the official ORACLE-L FAQ: http://www.orafaq.com

To unsubscribe send email to: oracle-l-request_at_freelists.org put 'unsubscribe' in the subject line.

--
Archives are at http://www.freelists.org/archives/oracle-l/
FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------

Received on Thu Aug 26 2004 - 02:18:26 CDT