[SunHELP] Tape drive problem

[Michael A. Turner]

	Hello all,
	
	we are trying to get the tape drive attached to our sunfire 280R
working and I am having problems. First the background info. The tape drive
is External in a 411 style case. It is a DDS4 tape drive, single tape. We
are running Solaris 8 on this system. The tape drive is raw device
/dev/rmt/0 . The drive has a brand new fuji DDS4 tape in it that came with
the system. All teh examples below are done once I am SU into root. The
cable to the device looks tight. Their is a storedge array but that is
connected to a different SCSI connector on the back of the sunfire. There is
no terminator on the Tape drives other connector and nothing is attached to
it.


	Now the symptoms.

mt -f /dev/rmt/0 status shows
HP DDS-4 DAT (Sun) tape drive:
   sense key(0x6)= Unit Attention   residual= 0   retries= 0
   file no= 0   block no= 0

mt -f /dev/rmt/0 rewind and mt -f /dev/rmt/0 erase seem to both work. The
drive operates as if it is working

tar has problems though. 

/users/mturner [11:31am]->tar cvf /dev/rmt/0 setpath
a setpath 1 tape blocks
/test [11:34am]->tar xvf /dev/rmt/0 setpath
tar: tape read error

on the Console I see

Aug 26 11:34:48 TeamWHRO scsi: WARNING: /pci at 8,700000/scsi @6,1/st at 4,0
(st18):
Aug 26 11:34:48 TeamWHRO         SCSI transport failed: reason
'unexpected_bus_free':giving up

the stats for file setpath are

-rwxrwxrwx   1 root     other        246 Jul 31 15:19 setpath

running the command again I get

/test [11:34am]->tar xvf /dev/rmt/0 setpath
tar: tape read error

and to the console

Aug 26 11:39:57 TeamWHRO scsi: WARNING: /pci at 8,700000/scsi at 6,1 (glm1):
Aug 26 11:39:57 TeamWHRO        gross error - not connected
Aug 26 11:39:57 TeamWHRO scsi: WARNING: /pci at 8,700000/scsi at 6,1 (glm1):
Aug 26 11:39:57 TeamWHRO        got SCSI bus reset
Aug 26 11:39:57 TeamWHRO scsi: WARNING: /pci at 8,700000/scsi at 6,1/st at 4,0
(st18):
Aug 26 11:39:57 TeamWHRO        SCSI transport failed: reason 'reset':
giving up

attempts on a larger file tend to go much more poorly

-rw-r--r--   1 root     other    9095680 Aug 26 11:14 Archive27546.tar

Taking an extremely long time to time out. Kill -9 will not stop the tar
process

/users/mturner [12:47pm]->tar cvf /dev/rmt/0 Archive27546.tar
a Archive27546.tar 17765 tape blocks
tar: write error: I/O error

With the same sort of SCSI error going to the console 

Aug 26 12:49:17 TeamWHRO scsi: WARNING: /pci at 8,700000/scsi at 6,1 (glm1):
Aug 26 12:49:17 TeamWHRO        got SCSI bus reset
Aug 26 12:49:17 TeamWHRO scsi: WARNING: /pci at 8,700000/scsi at 6,1/st at 4,0
(st18):
Aug 26 12:49:17 TeamWHRO        SCSI transport failed: reason 'reset':
giving up

Control C and Control Break do not stop the operation. The telnet window is
locked in the tar command.
again we see kill -9 failing

/users/mturner [12:51pm]->ps -A | grep tar
 28322 pts/4    0:00 tar
/users/mturner [12:57pm]->kill -9 28322
/users/mturner [12:58pm]->ps -A | grep tar
 28322 pts/4    0:00 tar

ten minutes later I get this at the console. I see killed at the telnet
window

Aug 26 12:59:26 TeamWHRO scsi: WARNING:
/pci at 8,700000/scsi at 6,1 (glm1):
Aug 26 12:59:26 TeamWHRO        Connected command timeout for Target 4.0
Aug 26 12:59:26 TeamWHRO scsi: WARNING: /pci at 8,700000/scsi at 6,1 (glm1):
Aug 26 12:59:26 TeamWHRO        got SCSI bus reset
Aug 26 12:59:26 TeamWHRO scsi: WARNING: /pci at 8,700000/scsi at 6,1/st at 4,0
(st18):
Aug 26 12:59:26 TeamWHRO        SCSI transport failed: reason 'timeout':
giving up

Googling so far has turned up ntohing. does anyone havea  clue to what is
causing this?


Michael A. Turner
Systems Engineer WHRO
michael.turner at whro.org
http://www.whro.org