There is no real "partition" for the 3494 like you are used to with the
3584.
Cartridges are mixed in there, you just set up separate category codes for
each host.

(It will make a little more sense if you go look at the 3494 Library Manager
console or web interface and list the tapes ) If your Z/OS has multiple
LPARs, it probably is already using multiple category codes.

Look at the TSM DEFINE LIBRARY command - when you specify 3494X, you also
specify SCRATCHCATEGORY and PRIVATECATEGORY codes.  Doesn't matter what they
are, as long as they are different from the codes already in use by z/OS.
(This is actually covered in the redbook, it's just such an anticlimax that
it's not terribly noticeable.)

You'll REALLY want to use different barcodes for your TSM tapes than for
z/OS.
You load tapes into the I/O door, and the 3494 files them away with an
"insert" category code (this is an "up for grabs" code).  When you do your
checkins, make sure that you specify a VOLRANGE so that you only grab your
own tapes.  As part of the checkin, the 3494 category code will get assigned
to YOUR tapes.  The 3494 will not let any other host use them, or let you
use tapes that has another host's category code.

That's all that's required for "partitioning".  Remember however, it
requires a different type of driver..

W


On Tue, Feb 17, 2009 at 11:19 AM, Nick Laflamme <dplaflamme AT gmail DOT 
com>wrote:

> Can anyone refer me to the parameters options for partitioning a 3494 ATL?
> It's easy to find that a 3494 can support two VTS partitions; I'm
> questioning how we add TSM on AIX to the existing z/OS non-VTS use and a
> VTS
> partition for z/OS.
>
> We have new 3592s to coexist with the z/OS's 3590s, so there's no question
> of sharing drives or volumes. I simply can't verify that we can set up a
> another partition for this new image.
>
> I have a copy of SG24-4409, the Redbook on partitioning 3494s. It doesn't
> address the question and, worse, it hasn't been updated in a decade.
>
> (I miss the 3584s I used at former sites. Sniff.) (Then again, someone said
> a single 3584-L23 is about $300,000, so I guess I understand why they
> didn't
> go that route right away.)
>
> Thanks,
> Nick
>