Measuring DiskSuite and Veritas performance hit
2007-12-24 21:46:00
I have been tasked with measuring the system hit taken by Veritas and
DiskSuite. We're considering upgrading to A5200's or A1000s from our
D1000 arrays, but we have to have justification.
Currently we're mirroring disks using disksuite. I use sdtperfmeter and
Disksuite statistics graphs for monitoring the systems usually, but what i'm
looking for is something more straight forward - for instance, instead of
disk I/O wait, how much of the processor time is being used by the md driver
to mirror the data thats coming in.
Currently we have to servers, for instance, Identical in every way except
that one does not use DiskSuite to mirror its drives. The one that does
not mirror is significantly faster than the one that does.
Can anyone suggest any methods of testing? I'd really appreciate it.
Thanks,
DP
_________________________________________________________________
MSN Photos is the easiest way to share and print your photos:
http://photos.msn.com/support/worldwide.aspx
Comments
Got something to say?
You must be logged in to post a comment.
