Re: How many simultaneous tape paths can a TSM server dr ive

Allen,
        Thanks for your reply.  Rephrasing my question, I suppose what I
really want to know is, does a session to a tape drive require more
resources from the TSM server than a session to a disk storage pool?  If
they are the same, then we have no problem.
        I just had never seen a 200 tape drive TSM server before.  EMC (who
makes the CDL) will support up to 512 virtual drives (depending on the
model), so support from the CDL is not the issue.  I just was concerned that
the TSM server itself might cave under the strain.  Perhaps I am worrying
for nothing. 


Best Regards,

John D. Schneider
Technology Consultant - Backup, Recovery, and Archive Practice
EMC² Corporation, 600 Emerson Road, Suite 400, St. Louis, MO 63141 
Phone: 314-989-3839 Cell: 314-225-9997 Email: Schneider_JohnD AT emc DOT com

-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
Allen S. Rout
Sent: Monday, November 07, 2005 11:08 PM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: [ADSM-L] How many simultaneous tape paths can a TSM server
drive

==> On Mon, 7 Nov 2005 16:08:04 -0500, John Schneider
<Schneider_JohnD AT EMC DOT COM> said:


> This will seem like an odd question, but can anybody tell me how to
> calculate the relative impact of tape drives on the performance of the TSM
> server?

[ ... ]

> But I have another customer who believes they need 200 virtual drives, and
> that all of them will be working at once.  They have that many
simultaneous
> client sessions going to disk pool, but they would like to drive straight
to
> CDL virtual tapes, and eliminate the disk pool.  Does anyone see a problem
> with that?


I would say ask the vendor of your virtual tape tech.  In theory, there's no
reason for the virtual tape behavior to be worse, performance wise, than a
similar number of writes to the underlying disk tech.  200 simultaneous
incrementals isn't out of whack for a medium sized installation, so you
ought
to be able to manage the raw I/O.

If (? Clariion ?)  blanches at the mount count, then you can have a few nice
rounds of Vendor Pinata while they explain why.  I'm sure some IBMers would
be
pleased to cough "SANergy" or some such in the background of that
conversation.


In fact, that was my first question, which I sat on until last: if you've
already got TSM in the equation, why would you prefer virtual tape on disk
to,
say, FILE devclasses?   I'm sure there are plenty of good reasons for such a
call, but that would be the tool I'd reach for first.




- Allen S. Rout