The communications problem was corrected in level 4.2.2.6.
Should have included that in the first message.
The problem I described below will be encountered when
you have enough drive usage between the sharing tsm's.
Good luck.

Philip A. James, Systems Software Specialist
Software Services Unit
Information Technology Services Division / Data Center
California Public Employees' Retirement System
Phone: (916) 326-3715
Fax:     (916) 326-3884
Email:  phil_james AT calpers.ca DOT gov

 -----Original Message-----
From: Jolliff, Dale [mailto:xjolliff AT TI DOT COM]
Sent: Wednesday, December 04, 2002 2:12 PM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: 3494 Partitioning or TSM Library Manager


Have them take a look at PMR 02224,519 - we just upgraded from 4.2.1.15 to
4.2.30 because of an identified "internal defect".

I don't have all the details in front of me, but at the 4.2.1.15 level there
is apparently some communications issues.


-----Original Message-----
From: James, Phil [mailto:Phil_James AT CALPERS.CA DOT GOV]
Sent: Wednesday, December 04, 2002 3:50 PM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: 3494 Partitioning or TSM Library Manager
Importance: Low


We have a similar problem with tape drive sharing.  IBM summarized the
problem very well below.


This was the response listed in the PMR from IBM.

Researched problem and found that the root cause of this problem is the
same as that for pmr 04108,bre.
..
The dismount command issues a dismount request the the drive, we make
sure that the drive receives the command and we request a response.
If we do not get an immediate response we retry the dismount 5X.
..
If after the 5th retry we still cannot communicate with the drive
we issue an IO error and mark the drive offline.
..
In the event of library sharing and/or library fail-over a situation
can occur where the drive is not available to the TSM server that
issued the dismount command; in library fail-over the library is
unavailable for a period of time that exceeds the amount of time it
takes for the 5 retries (approximately 5 minutes), if the event of
another server taking control over that drive (since we immediately
release the drive after issuing the dismount request to the library)
the TSM server that issued the request can no longer communicate with
that drive and if this exceeds the retry period then we would mark
the drive offline.
..
While this is not preferred behavior it is not necessarily a defect.
..
A DCR will have to be opened to request that in the case of library
sharing and library fail-over the dismount request goes into indefinite
polling status instead of using a retry period.
..

We are currently having our IBM Marketing Rep. open the Design Change
Request (DCR). for this feature defect.

If anyone is experiencing this problem would you please have your
IBM Marketing Rep. open a DCR for you.  The more DCR's that are opened the
sooner the not preferred behavior can be corrected.
You can refer to our PMR # 32181.  All help will be appreciated.

Other than this the TSM Library Manager feature has been working fine.

Philip A. James, Systems Software Specialist
Software Services Unit
Information Technology Services Division / Data Center
California Public Employees' Retirement System
Phone: (916) 326-3715
Fax:     (916) 326-3884
Email:  phil_james AT calpers.ca DOT gov



-----Original Message-----
From: Davidson, Becky [ mailto:Becky.Davidson AT SLBG DOT COM
<mailto:Becky.Davidson AT SLBG DOT COM> ]
Sent: Wednesday, December 04, 2002 12:27 PM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: 3494 Partitioning or TSM Library Manager


I have implemented the TSM Library Manager feature.  I have 14 drives shared
among 4 TSM servers.  When I tried the categories and use whatever was
available if tsmsvr1 tried for drive 1 a couple of times and tsmsvr2 was
using it then tsmsvr1 marked it offline.  It made for a huge mess.  Now I
definitely have to keep tsmsvr2 (the server manager) up all the time unless
I am booting the other servers.  Occasionally I have had a check out for
tsmsvr1 show successful yet the tapes don't come out because there were not
enough drives.  occasionally I have also had where you have to cycle the tsm
software because there seems to be some communication problems.  Generally
that was when the operators didn't reply to the checkout and the checkout
timed out.  If I had a choice of whether to go back or not I wouldn't.  This
made things so much easier in the long run.  My tsm server are running with
one on an S85, two on a 6M1, and one on an S85 with all of them running aix
4.3.3 and after this weekend all of them running tsm 5.1  (two are still tsm
4.1.2)

-----Original Message-----
From: Joshua Bassi [ mailto:jbassi AT IHWY DOT COM <mailto:jbassi AT IHWY DOT 
COM> ]
Sent: Wednesday, December 04, 2002 10:19 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: 3494 Partitioning or TSM Library Manager


All,

I am deploying 3 new TSM server on a single p690 LPAR.  There is a 3494
with 32-tape drives attached to the system.  I have a choice to make
whether to use the old 3494 category scheme to separate my 3 TSM servers
or should I use the TSM Library Manger feature to control access to all
32 tape drives?

First of all, how many customers are using the Library Manager feature
of TSM?  What kind of caveats are there in deploying this feature?  How
well does the automatic tape drive allocation work?  Thanks in advance.


--
Joshua S. Bassi
IBM Certified - AIX 4/5L, SAN, Shark
Tivoli Certified Consultant -ADSM/TSM
eServer Systems Expert -pSeries HACMP

AIX, HACMP, Storage, TSM Consultant
Cell (831) 595-3962
jbassi AT ihwy DOT com