Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Mailing Lists -> Oracle-L -> RE: Optimal Disk Layout??

RE: Optimal Disk Layout??

From: Christopher Spence <cspence_at_FuelSpot.com>
Date: Fri, 21 Sep 2001 07:24:38 -0700
Message-ID: <F001.00394ABA.20010921073540@fatcity.com>

  1. depends on the stripe width, data may be on one or many drives, parity will always get written to all.
  2. I believe your performance will be better; you have 6 drives for data, and 4 for archive/redo. That equates to 12 drives over the previous 8. However, you were using all eight for everything.

I don't believe there is any purpose to mirror the four drives, I like to use Raid 0 or single drives for redo, and just do multiple members and archive destinations. Redo/Archive is high write activity, Raid 1 takes a small hit on write, and only benefits read performance. You can get the same redundancy using duplex destination and multiple members.

Having redo logs and archive logs separate will prove to be beneficial.

All and all, I think you will come up with a better configuration performance-wise.

Make sure the A1000 has the 80Mb cache, and has the latest firmware, as it there are a lot of buggy A1000's out there that cause corruption if they are saturated.

"Do not criticize someone until you walked a mile in their shoes, that way when you criticize them, you are a mile a way and have their shoes."

Christopher R. Spence
Oracle DBA
Phone: (978) 322-5744
Fax: (707) 885-2275

Fuelspot
73 Princeton Street
North, Chelmsford 01863  

-----Original Message-----
Sent: Friday, September 21, 2001 10:15 AM To: Multiple recipients of list ORACLE-L

Hi,

I'm encuntering a relatively high iowait percentage when my hot backups are running. The platform is Sun
(E420 2cpu running SunOS 5.7) with an A1000 disk array
(8 9Gb drives, hardware Raid-5). The array is one
volume and all DB components (redo, archive, data, index, system, etc.) are on the same volume except for the binaries. The database is not large, only ~8Gb in size and the transactional volume is not much either. However, when the backups run, the iowait according to 'top' hovers between 50-70% which causes our application to time-out via Web Logic 6.0.

The developers can't explain why the timeouts are occurring (WLS 6.0 is a new upgrade from 5.1). The SysAdmin isn't much help either.

I have an opportunity to rebuild the database on another machine and use RAID 1+0 -- the thought being that we are choking ourselves with Raid-5 when the hot backups are performing the cp's (copies) and then the files are compressed.

My first question is, how is the data distributed across the drives in my Raid-5 configuration? Is each disk being filled contiguously in series or is the data being spread around in a pseudo-striping manner?

My second question/dilema is, the new array (another A1000) will have 6 18Gb drives and with Raid 1+0 that shrinks to 3 drives of usable space for everything except redo and archive. 4 9Gb drives will be added in two mirrored sets, one for redo and the other for archive. I'm afraid that I will see worse I/O performance with the new array because it has so fewer physical drives, thus eliminating the benefit of not having to write the parity info.

Do you concur? Knowing the two arrays I have to work with, which would be the better configuration?

Any suggestions, recommendations would VERY much be appreciated.

Thanks again for the feedback.
-w



Terrorist Attacks on U.S. - How can you help? Donate cash, emergency relief information http://dailynews.yahoo.com/fc/US/Emergency_Information/
-- 
Please see the official ORACLE-L FAQ: http://www.orafaq.com
-- 
Author: Walter K
  INET: alden14004_at_yahoo.com

Fat City Network Services    -- (858) 538-5051  FAX: (858) 538-5051
San Diego, California        -- Public Internet access / Mailing Lists
--------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L

(or the name of mailing list you want to be removed from). You may
also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: Christopher Spence INET: cspence_at_FuelSpot.com Fat City Network Services -- (858) 538-5051 FAX: (858) 538-5051 San Diego, California -- Public Internet access / Mailing Lists -------------------------------------------------------------------- To REMOVE yourself from this mailing list, send an E-Mail message to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from). You may
also send the HELP command for other information (like subscribing).
Received on Fri Sep 21 2001 - 09:24:38 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US