ADSM-L

Re: Defining TSM Disk pools on ESS wisdom wanted

2003-03-27 20:54:15
Subject: Re: Defining TSM Disk pools on ESS wisdom wanted
From: Alex Paschal <AlexPaschal AT FREIGHTLINER DOT COM>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Thu, 27 Mar 2003 14:56:22 -0800
To expand on Daniel's answer, you would want each RANK to be written to by
only one concurrent i/o thread. So, if several LUNs are on the same rank,
you might want to LVM them into a large volume, then define it to TSM.
Otherwise, if 2 LUNs are on the same rank, then when TSM is writing to both
volumes, you'll still get head contention.  How much that'll actually impact
you, well, that's up in the air also.  You may want to test.

Also, I'm not so sure JFS is the best way to go.  As near as I can tell,
it's a 6-or-half-dozen issue.  Personal preference seems to be the major
differentiation.  I've seen people claim that JFS is faster, I've seen
people claim RAW is faster, so I'd say it's pretty much up to you.  You may
want to test both in your environment.

Alex Paschal
Freightliner, LLC
(503) 745-6850 phone/vmail

-----Original Message-----
From: Daniel Sparrman [mailto:daniel.sparrman AT EXIST DOT SE]
Sent: Thursday, March 27, 2003 1:25 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: Defining TSM Disk pools on ESS wisdom wanted


Hi

I would suggest having as much volumes as possible. This is because AIX 
treats every volume as a PV, and one PV can only do one write or read at a 
time.

So, using 24GB volumes is probably the best way to go.  Also, JFS 
formatted volumes is the way to go.

Best Regards

Daniel Sparrman
-----------------------------------
Daniel Sparrman
Exist i Stockholm AB
Propellervägen 6B
183 62 TÄBY
Växel: 08 - 754 98 00
Mobil: 070 - 399 27 51




"Anthonijsz, Marcel M SITI-ITDGE13" <Marcel.Anthonijsz AT SHELL DOT COM>
Sent by: "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
2003-03-27 10:18
Please respond to "ADSM: Dist Stor Manager"
 
        To:     ADSM-L AT VM.MARIST DOT EDU
        cc: 
        Subject:        Defining TSM Disk pools on ESS wisdom wanted


Hi,

We are redefining our TSM disk pools on IBM ESS storage. We got 39 x 24 GB
LUNs from our storage collegues. Total 936GB.
LUNs are spread across the whole ESS. Now what is the best way, TSM wise 
and
performance wise to define these disks to TSM?

Make the volumes as large as possible?
or define a volume per LUN, so I get 39 volumes of each 24GB?
Use raw volumes or JFS formatted?

Almost forgot: we're running TSM 4.2.1.9 on AIX 4.3.3 ML10 (yes I do
know....All going out of support soon).
ESS is FC connected.

Thanks all for your eternal wisdom,


Marcel Anthonijsz
Central Data Storage Manager (a.k.a. storman)
Shell Information Technology International B.V.
PO Box 1027, 2260 BA  Leidschendam, The Netherlands

Tel: +31-70 303 4984
Mob: +31-6 24 23 6522
Email: Marcel.Anthonijsz.at.shell.com

<Prev in Thread] Current Thread [Next in Thread>