The original problem was that the gripper would place a tape in the drive,
but not far enough in to be loaded. The 3494 console showed the mount
request in the queue. ADSM showed a mount request. The tape drive in
particular showed "busy" when I used the mtlib command against it.
Otherwise the library was fully functional. All other tape requests were
being processed. I left the process running for 72 hours (decided to wait
till the weekend was over and I was back in the office), and ADSM was still
patiently waiting for the tape mount.
> -----Original Message-----
> From: decook AT amoco DOT com [SMTP:decook AT amoco DOT com]
> Sent: Wednesday, March 17, 1999 11:13 AM
> To: HiltonT AT tce DOT com
> Subject: Re: mountwait not being used
>
> Now I didn't catch the message on what was being seen with the
> mountwait "stuff" but...
>
> Seeing how it was an atl failure, if ADSM can't talk to the library
> or
> the library knows its "broke, don't work" it might handle things
> differently.
>
> Have you used the "mtlib" command ?
> (mtlib -? will yield a help screen)
> useful options are
> mtlib -l/dev/lmcp# -qL
> (where # is 0-9 found by dsmadmc> q libr or aix> lsdev -Cc tape)
> query library, shows the state of the library, if the input & output
> stations are empty (convenience I/O station), accessor, gripper, &
> vision status... also down at the bottom lists # of 3490 & 3590
> cleaning cycles left on the cleaning tapes.
> mtlib -l/dev/lmcp# -qI
> query the inventory, to compare to a dsmadmc> q libvol in order to
> look for "lost" volumes
>
> then within adsm to check on the library you may use the
> dsmadmc> show library
>
> used to mainly be good to see the "busy" and "online" stats of the
> library and its drives but now-a-days a "q drive" shows if and when a
>
> drive died. Still good to "double check" things.
>
> Oh, here is just a little helpful trick, if you have a means of
> issuing a page (or use it to send you mail or trigger some sort of
> notification) this runs every hour... with 6 ATL's and 13 adsm
> servers
> a bad day can really be a BAD DAY!
>
> later
> Dwight
> ;-)
>
> #!/bin/ksh
> # first reset counter file to limit pages
> if [ $(date +%H) -eq 8 ] ; then
> cat /dev/null > $HOME/bin/librck.out
> fi
> for LMCPX in $( lsdev -Cc tape | grep ^lmcp | cut -d' ' -f1 ) ; do
> LIBINFO=$(mtlib -l/dev/$LMCPX -qL | cut -c30- | egrep
> '(Paused|Degraded|Inter
> vention)')
> if [ $? -eq 0 ] ; then
> echo $(hostname) $(date +%D,%T) $LMCPX $LIBINFO >>
> $HOME/bin/librck.out
> if [ $(cat $HOME/bin/librck.out|wc -l) -lt 3 ] ; then
> rsh pagesrv autopage adsm $(hostname) $(date +%D,%T) $LMCPX
> $LIBINFO
> # echo $(hostname) $(date +%D,%T) $LMCPX $LIBINFO
> fi
> fi
> done
> exit
>
>
>
>
> ______________________________ Reply Separator
> _________________________________
> Subject: mountwait not being used
> Author: HiltonT (HiltonT AT TCE DOT COM) at unix,mime
> Date: 3/17/99 9:02 AM
>
>
> On Monday, I mentioned that adsm wasn't following the mountwait parameter
> when I was having problems with my 3494. Well, the CE was called
> yesterday
> and shortly after the call the 3494 failed completely. It ended up being
> a
> problem with the gripper and a part was replaced.
>
> I talked to ADSM support this morning, and they said that the mount
> request
> should have timed out. They agree there's something that's not working
> correctly in ADSM. If it happens again, level 2 will provide traces that
> should help pinpoint the problem. So, if it happens to anyone else on
> this
> list, please report it to IBM support and let the rest of us know what
> happened.
>
> Thanks to everyone who commented, offered suggestions, or just
> commiserated.
>
> Tina Hilton
> Thomson Consumer Electronics
> Systems Software
> 317-587-4292
> hiltont AT tce DOT com
|