ADSM-L

Re: [ADSM-L] question on 3494/3592

2007-11-01 18:27:46
Subject: Re: [ADSM-L] question on 3494/3592
From: Richard Sims <rbs AT BU DOT EDU>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Thu, 1 Nov 2007 18:27:01 -0400
On Nov 1, 2007, at 5:40 PM, cc52 wrote:

We have been trying to set up two J1a drives in a 3494 for a Windows
instance of TSM.  To test out the drives we have been able to mount
a new 3592 tape with mtlib.  We are then able to run NTUTIL and run
some
tests on the drive.  After closing it we try to use mtlib to dismount
tape and put it back in the 3494 slots.  At this point the tape is
put into UNLOAD state sitting in the 3592 tape drive and an
int req pops up on the 3494 indicating the tape has been dropped.

This type of test works fine on our AIX 3592's. We are using the
latest
drivers on windows.  We're about ready to put TSM on the windows box
(2003) and see if it interacts any different than mtlib/nutuil tools
do.


Any ideas on what could be messing up the demount function on the
kibrary?

Ron -

What you report points to a defective gripper situation, which needs
an IBM CE to replace parts.  3494 robotic tape handling is a function
of native, embedded logic in the 3494 library and has nothing to do
with host drivers.

We had much the same with 3592 tapes but not with the 3590s also in
the AIX-controlled library: the robot would return a tape to the cell,
but *sometimes* leave it partly sticking out of some cells - and then
collided with it, causing the library to shut down.  After much
attempts to rectify the situation with unsuccessful measures, the CE
finally replaced both "hands" on the actuator, and that put an end to
the problems.  The 3592 tapes are somewhat different than the 3590s,
which seems to be the root cause, such that the original "hands" on
the robot weren't quite fully capable of grasping the 3592s correctly.

    Richard Sims

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