SUMMARY: Performance numbers?

From: James Benson <jbenson_at_s5t.com>
Date: Fri Nov 09 2001 - 11:26:18 EST
Many thanks to all who responded; there are too many to list.

Overwhelming consensus among respondents was that software RAID 5 was a big
hinderance to the environment's performance.  I have recommended to the
client that they add another SCSI channel via another SCSI card, purchase 2
additional disks for the D1000, and switch to RAID 1+0.

Although I was not given any performance numbers, Sun engineer
(guess)timated that increased performance, depending on users, apps, etc,
could be in the 20-40% range.

Thanks again for all the help for a first-time poster.

Regards,

James Benson
Systems Engineer
S5T/BCG, a division of Pomeroy
919.677.8514  x2005
jbenson@s5t.com

-----Original Message-----
From: James Benson [mailto:jbenson@s5t.com]
Sent: Thursday, November 01, 2001 11:18 AM
To: sunmanagers@sunmanagers.org
Subject: Performance numbers?

All,

I have come across a request for real-time performance numbers, if they are
available (hopefully, someone on the list can point me in the right
direction).

We have a client with an E220R, with a D1000 attached, single-bus SCSI.
They are running one instance of Oracle and are commenting that the
performance is somewhat slow (no surprise, really).  Right now our only
option is to add another SCSI connection to increase performance, but I have
been asked to show any numbers available to show exactly how much
performance difference we're looking at.

Configuration:
E220R, (2) 450MHz CPU, 2GB RAM
2 internal disks, mirrored using SDS
D1000 attached via dual diff SCSI
L9 attached to other SCSI port on the card
D1000 disks: (4) 18GB, RAID 5 using SDS

Would welcome other recommendations, but keep in mind I will be asked for
real-time perf numbers to back up any suggestions.

Thanks in advance for any help.

_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
Received on Fri Nov 9 10:26:15 2001

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:42:27 EST