Ooops... I would beg to differ with the "expert" !
Now as load fluctuates this will shift from being important to not so
important but...
TSM really works good at spreading its load across all defined logical
devices.
If you have each 8 GB physical device split into 4 - 2GB logical devices
then say you have 4 physical devices to yield 16 logical devices (files
defined to tsm)
now if you have 8 client sessions initiate, tsm will spread that inbound
data across its devices (files) you stand the chance of the 8 inbound
sessions going to only 2 physical devices and having 2 physical devices
idle.  (doesn't seem like the kind'a solution I'd want)
If you were to just define 1 logical device per physical device yielding 4
tsm devices (files) then tsm would spread the load out to more evenly across
all physical devices...
In greater detail you'll notice work is laid out across dasd almost by the
individual packets being received...
but if you put more than one adsm logical device per physical device you
stand the chance of over working one physical device while another stands
idle...
Hey ! should I be charging consulting fees here  ?   ;-)

Dwight



> ----------
> From:         Davidson, Becky[SMTP:DAVIDSBX AT EGR DOT COM]
> Reply To:     ADSM: Dist Stor Manager
> Sent:         Wednesday, July 26, 2000 2:11 PM
> To:   ADSM-L AT VM.MARIST DOT EDU
> Subject:      Re: Hard Disk
>
> We simply def additional disk volumes using
> add vol to volume group
> either make logical volume or format a volume
> define volume stgpool filename formatsize=xxx
>
> We had an expert come in and he reccommended partitioning our disks into
> multiple segments and having several segments like on an 8 gig disk having
> 4
> 2gig volumes of storage.  That was for performance.  You probably won't
> need
> to do that with the shark depending on how you set it up.  Please let me
> know how that works performance wise with the shark because we just got
> one
> and will be doing that to a certain degree in the near future.  Another
> performance boost that we got was on my disk pool I don't start migration
> until 75% and then I go down to 1.  This has more time to dump to disk
> with
> less migration going.  It usually keeps up.
>
> -----Original Message-----
> From: Shekhar Dhotre [mailto:Shekhar.Dhotre.B AT BAYER DOT COM]
> Sent: Wednesday, July 26, 2000 1:47 PM
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: Hard Disk
>
>
> Hi all
>  Currentlly i am having 2gb , as a primary stg. pool,, and we are backing
> up
> more than 2GB per night, i have set low and high threshold values so that
> it
> can start migration , as the data touches to High threshold point (30-70).
> but
> instaed of that i want to allocate more disk space to primary stg pool.
> which i want to  make 20 GB.   Our  TSM 3.7.2 is on S-70 (AIX 4.3.3.)
> we are using IBM E20 SHARK  as a sorage unit and i want to allocate disk
> from
> this unit .
> i know AIX  and shark  well , so i just want to know steps involed
> allocating
> more disk space for , primary stg. pool.
>
> Thanks in Advance .
> Shekhar
>