Re: Veritas -- do # of spindles really matter ??
From: Andrew Garman <algarman_at_lucent.com>
Date: Sat, 15 Jan 2000 18:28:53 GMT
Message-ID: <20000115.18280800_at_mokey.>
> The question:
> Which is faster -- (a) the current 2-spindle architecture, or (b) if 5
> separate filesystems were explicitly-defined on their own dedicated
disk?
> Keep in mind both (a) and (b) would take advantage of the same
Veritas/RAID
> configuration (striping and mirroring).
Received on Sat Jan 15 2000 - 19:28:53 CET
Date: Sat, 15 Jan 2000 18:28:53 GMT
Message-ID: <20000115.18280800_at_mokey.>
I'd suggest (b) as I don't really understand what was being done in
(a). My
best guess is that you had five disks with multiple volumes defined on
them.
That's generally not the best way of using Volume Manager. If at all
possible,
to attain maximum performance you should have only a single volume
defined per
disk. With large disks, this tends to be difficult to arrange.
Having multiple
volumes on the disk will increase the likelihood of disk contention.
This can
be controlled if you configure the different volumes with workloads
that do not
overlap.
If you can, test the two configurations and see which one performs best.
Best of Luck,
Andrew Garman
>>>>>>>>>>>>>>>>>> Original Message <<<<<<<<<<<<<<<<<<On 1/13/00, 6:54:22 PM, "jawa" <Jim.Wadas_at_motorola.com> wrote regarding Veritas -- do # of spindles really matter ??:
> The question:
> Which is faster -- (a) the current 2-spindle architecture, or (b) if 5
> separate filesystems were explicitly-defined on their own dedicated
disk?
> Keep in mind both (a) and (b) would take advantage of the same
Veritas/RAID
> configuration (striping and mirroring).
Received on Sat Jan 15 2000 - 19:28:53 CET