Hmm, we saw a similar problem in our 3494 library when we upgraded
to the 3590E drives. Perhaps this is like your problem, maybe not...
It was hard to catch it in the act, because it was intermittent, but
when we finally did, we saw that the gripper tried to load the tape in the
drive, but it never successfully loaded. The tape was just millimeters from
being loaded, and if we paused the library and tapped the tape, it would
load.
Some of the symptoms we saw were:
1. "q sess f=d" and "q proc" would show very long "waiting for mount
of volume..." times.
2. On the 3494 interface, we could see that it thought the tapes
being requested were loaded, but ADSM never mounted them.
3. Eventually the drive would time out an and become unavailable (I
think, my memory is a little fuzzy on this point).
Our CE tried many things:
1. Reteach library about the location of the drives ( no change).
2. Load the latest drivers (no change).
3. Put new belts on the grippers (no change).
What he eventually determined was that on all 3590 drives, there is
a spring loaded flap in the mouth of the drive that holds the tape so it
won't fall out while waiting for the gripper to retrieve the tape. On the
3590E drives, instead of one spring holding the flap, there are 2 springs.
If you manually try to pull a tape out of a drive, you will notice that they
are alot harder to pull out on the 3590e drives than the 3590B drives.
The higher tension of the 2 springs (and the increased friction)
kept the tape from sliding all the way into the drive and caused our
problem. Our CE removed the extra spring from the 3590E drives and we have
not had the problem since.
Ben
> -----Original Message-----
> From: Forrester, Benjamin S. [mailto:ForresterB AT MISSOURI DOT EDU]
> Sent: Wednesday, November 17, 1999 10:30 AM
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: Re: Storage media inaccessible in a 3494 library
>
>
> That is what is particularly strange about the error, there
> can be at least
> two mount points available, sometimes more in the library
> when this happens.
> There is nothing physically wrong with the drives and you can see the
> cartrige sitting in its slot in the library. Moving data on
> other carts work
> just fine, but we get the error with just a select number of
> volumes. I've
> tried the moves at different times of the day, when no other
> processes are
> running but the error is still the same. Any more ideas are
> very welcome.
>
> Ben Forrester
>
> -----Original Message-----
> From: Vraspir, Mary J. [mailto:MJVraspir AT WEST DOT COM]
> Sent: Wednesday, November 17, 1999 11:05 AM
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: Re: Storage media inaccessible in a 3494 library
>
>
> Move date will fail if there is no empty drives. It is also
> one of the
> first processes to be preempted. Make sure no one is using
> the drives or
> there is not tapes mounted in any drives you are expecting to
> use. You can
> issue dismount vol xxx if necessary.
> MV
>
> -----Original Message-----
> From: Forrester, Benjamin S. [SMTP:ForresterB AT MISSOURI DOT EDU]
> Sent: Wednesday, November 17, 1999 10:22 AM
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: Storage media inaccessible in a 3494 library
>
> I am trying to move data from some suspect volumes in
> our library,
> but the
> move datas consistently fail. The accounting log gives me the
> following
> errors:
>
> 11/17/99 09:55:11 ANR8447E No drives are currently
> available in
> library
> LIB3494A.
> 11/17/99 09:55:11 ANR1401W Mount request denied
> for volume A00088
> - mount
> failed.
> 11/17/99 09:56:25 ANR1144W Move data process terminated for
> volume A00088
> -
> storage media inaccessible.
> 11/17/99 09:56:25 ANR0985I Process 126 for MOVE
> DATA running in
> the
> BACKGROUND completed with
> completion state
> FAILURE at
> 09:56:25.
> The problem is that there are available 3590E drives
> available when
> I issue
> the command, and issuing the move data command on a
> different volume
> works
> perfectly. Has anyone seen this before or have any
> ideas as to what
> is
> wrong? Is the data or the cart so corrupt that it
> cannot be mounted?
> Any
> comments or suggestions would be very helpful. We are
> at ADSM level
> 3.1.2.40
> and have the latest microcode for the 3590E drives.
> The following
> information is the detailed information on the volume
> I was trying
> to move.
>
> Thanks,
>
> Ben Forrester
> System Administrator
> University of Missouri - Columbia
>
> Volume Name: A00088
> Storage Pool Name: BACKUPCART
> Device Class Name: 3590
> Estimated Capacity (MB): 10,240.0
> Pct Util: 15.6
> Volume Status: Filling
> Access: Read/Write
> Pct. Reclaimable Space: 29.9
> Scratch Volume?: Yes
> In Error State?: No
> Number of Writable Sides: 1
> Number of Times Mounted: 13
> Write Pass Number: 1
> Approx. Date Last Written: 10/07/99 05:18:23
> Approx. Date Last Read: 10/08/99 10:03:43
> Date Became Pending:
> Number of Write Errors: 0
> Number of Read Errors: 0
> Volume Location:
> Last Update by (administrator): DILLB
> Last Update Date/Time: 11/16/99 23:49:28
>
|
