Geoffrey,
I've seen this several times on my 3494. One of the drives is giving
"false" status of availability back to TSM. You should be able to
determine the drive by examining the TSM activity log (q actlog) and
looking for which drive has NOT had any successful mount requests. The
cause (at least in our case) is a tape volume gets "pinned" in the mount
queue in the ATL. This can happen when there is some kind of rare
circumstance involving a mount/dismount failure (or perhaps after a CE has
done maintenance??). Normally, the ATL and TSM handle such events quite
nicely, but every now and then, the two get out of sync. TSM thinks the
drive is free, but there's one or more entries pinned in the mount queue in
the ATL. So, TSM schedules operations on the ATL based on an incorrect
assessment of drive availability and the ANR8447E messages start to show
up. In our case, we were flooded with such messages because TSM apparently
isn't smart to enough to stop trying after so many failures. Here's how
you check for - and solve - this problem: To check the ATL, select
"queues", then "mount queue". Normally the screen should be empty (i.e.,
nothing in the queue). If you see something, and it persists for a while
(e.g., more than a minute), then this is pretty good evidence of a pinned
queue request on the indicated device (e.g., device 101). This device
would/should be the same as the drive you determined by examining the TSM
activity log as described earlier. This pinned request will NEVER go away
on its own unless you take corrective action: Go into service mode
(password required) on the ATL. Select the "availability" menu, and then
select "set physical tape subsystem availability". Click on the "change
availability" button at the bottom. Locate the checkbox representing the
drive involved (e.g., drive 101). Uncheck the drive and click "change
availability". Then re-check the checkbox and again click "change
availability". Then, go back and check the mount queue which should now be
totally empty. At this point, your ATL should be working properly again.
That said, we also experienced this problem on another occasion that was
traced down to how the drive was physically cabled, and only after
switching the cables on that and another drive, and redefining the drives
to TSM were we able to clear the problem. I hope this information helps.
Bob Brazner
Johnson Controls, Inc.
(414) 524-2570
|
