bradd@gssc.gss.com (Brad Davis) (03/25/91)
I have a suspected compatiblity problem between two SCSI devices that I
would appreciate your advice on. Let me describe the problem in a general way
before I get into the specifics.
I have a SCSI host adaptor that is connected to a non-embedded SCSI tape drive.
By which I mean the host adapter is plugged into the host machine and the
tape drive is actually in two parts: a controller card that talks SCSI on
the one side and then turns around and talks QIC-24 (I think) to the tape
drive. The controller is housed in a metal box the size of a half height
drive and attached to the bottom of the half height tape drive, and the two of
them are screwed together to form sort of a full height SCSI tape drive.
The tape backup software recognizes the drive, and can read and write
small things like volume labels just fine, so it looks like the cables
are correct.
The problem is that as soon as I backup anything at all large, the data
will not be written (read?) correctly for more than 74k bytes at a time.
Out of a 55 megabyte backup, I had 832 blocks fail verification, which I
think means I had some good sections and some bad sections. The tape
cartridges are brand new, so I kind of doubt that the media is at fault.
Now for the specifics of the hardware involved:
Adaptec AHA-1540 ("ASSY 411306-01 / Rev B")
(Irq 11, DMA 5, ROM at DC00, Port 330)
Emulex "assy MT 0210403 Rev K"
60Mb Archive tape drive, model 5945C
Dysan 600 tape cartridges (600 ft, Hci 5550, 12500ftpi)
I stripped the host system down as far as possible to simplify things:
25 MHz 386 clone (C&T chipset, AMI bios) with 8Mb on motherboard
MFM hard/floppy disk controller driving
80 Mb Seagate 4096 hard disk (one DOS 3.31 partition)
1.2Mb & 1.44Mb floppy disk drives
16-bit VGA (Paradise chipset)
MS DOS 3.31
SY-TOS Version 3.03 AD
No TSRs or device drivers except for ASPI4DOS.SYS (needed by SYTOS)
And now some speculation: I'm guessing that something in the sytem cannot
keep up, and periodically the drive will want to ship (or receive?) data only
to find that there is none ready for it. The drive has to back up and wait
and then continue when more data is ready, but somewhere along the line the
handshaking gets fouled up and some junk gets written (or read). The physical
motion of the tape is very much a two steps forward/ one step back kind of
thing which tends to support the theory.
Can anybody tell me how to get the Adaptec, Emulex and Archive to work
together properly, or is this a lost cause?
Please email me (bradd@gssc) if possible (I'll get it quicker that way) or
post if you prefer. I will post a summary if there appears to be interest.
Thanks for your time!
Brad
_________
Random drivel from the keyboard of: | |.|
Brad Davis, GSS Inc, Beaverton OR | bradd |:|
bradd@gssc (503) 671-8431 ___|_______|_|___ _________
Disclaimer: The boss disavows |[]gssc| _-_ _-_ | @|#########|
all knowledge of my actions. |______|__________|@@@@ +---------+