Thanks to,
wfong_at_synacom.com
alan_at_nabeth.cxo.dec.com
Zack Ellis,
Dr. Tom Blinn,
****************************************************************************
****************************
ORIGINAL MESSAGE
On an AS1200 running DU4.0d backups are done via cron with vdump.
Last night it seems as if the tape ran out of space and now the tape is
stuck in the machine.
I cannot seem to find the process that is running which is holding the tape
in.
The drive is a TLZ10 and the tape was a 125m 4mm tape.
what should i look for ?
****************************************************************************
****************************
To free up the tape in the end i just rebooted the machine. This reset the
drive and allowed me to eject the tape.
I got a few more options i could have tried before i rebooted the machine
which I will try next time. !
1)
Did you consider the possibility that what is keeping the tape stuck in the
drive is a drive fault of some kind? The error message from vdump doesn't
say that it ran out of tape, it says there was a write error, typically an
end of tape condition is NOT reported as an I/O error.
If you can power cycle the tape drive, it may clear the condition and let
you get the tape out.
As it is an internal tape drive i was unable to recycle the power on the
tape drive alone.
2)
I have seen similar problems with my 2100, I issued a
mt -f /dev/rmt0h status
command to find out if it's still responding, then you can do a
mt -f /dev/rmt0h online to "wake it up"
This would not work as the system was replying to me saying no such device
i.e i was unable to get the status.
3)
An I/O error is unlikely to be EOT, so check the error log to see
if there really was an error around that time. You should use
the dia command (DECevent), but can use uerf(8). If you use uerf
be sure to use the option "-o full" so you get the full listing.
If the drive doesn't respond to pushing its eject button, then
you do seem to have a hardware problem with the drive.
The "Not a typewriter" message is a consequence of vdump trying
to open /dev/tty for the process doing the backup. Since cron
doesn't give processes a controlling tty it is understandable
that trying to open it will fail. This is just a limitation
of running backups via cron; the utilities expect a controlling
terminal through which to prompt for user input.
4)
I have seen this problem with an AS2100. Do a ps -ef and kill stray dumps,
or the shell that was called by your cron job. I also have had to look at
"ipcs" and do a few 'ipcrm" to kill off the wait flags. ipcrm must be done
very carefully, or you will kill off or hang running processes. Or, you
could try scu to reset your tape drive. Try:
scu -f /dev/rmt[nn] reset device
Received this after the reboot.
Thanks again,
Darrell King
Systems Manager
+353 1 6142430
darrellk_at_apion-ti.com
087-2850314
Received on Thu Apr 22 1999 - 16:30:07 NZST