by SCSI disk drive write errors using Linux with an UltraStor 24F
I provide the following information in the hope that someone knows of
or has experienced my problem and may offer a suggestion. I don't
know enough about writing drivers to try to fix this and I'd really
like to pursue using Linux.
Specifically, I am trying to use:
Linux 1.83 on Ultrastor 24F EISA SCSI Adapter w/Seagate ST41601N 1.6
giga. drive. I am consistently having problems whenever I try to
build a new kernel on the Seagate or copy any large file to it. For
example, trying to copy the kernel sources .tar.z file, or copying a
group of 100k to 700k files. Numerous messages appear about the SCSI
commands being retried due to errors and the process does not complete
successfully.
Some actual examples:
scsi0 : resetting for second half of retries
US14F : reset: called
SCSI disk error : host 0 id 0 lun 0 return code=27070000
scsidisk I/O error: dev 0803, sector 83442
.
. more of the same 1st 2 lines
.
Ux4F multiple commands completed
or
SCSI disk error : host 0 id 0 lun 0 return code=28000002
extra data not valid Current error sd803 : sense key Unit Attention
Additional sense indicates Power on, reset, or bus device reset
occurred
scsidisk I/O error: dev 0803, sector 81934
In the case of copying many files, several seem to copy without error,
others have a lot of variety in the level of corruption. Attempts to
make a new kernel, however, always fail. Sometimes I get the
"kernel panic" message along the way.
Since I had similar errors repeatedly trying to install the system
initially, I had to swap the UltraStor for a borrowed Adaptec 1740
with which I had no problems. I then put the UltraStor back in the
system and I can now boot Linux successfully, start X-windows and run
the suite of demo software that came with the Yggdrasil Fall-'94
distribution without error, but I am still not able to do any kernel
rebuilds or copy large files to the drive without these errors.
Question: have I created additional problems for myself by doing the
initial Linux install using the AHA1740 controller and now trying to
use the drive with the UltraStor?
My configuration is an Amtron DX-8100C EISA-VESA-ISA motherboard with
Intel 486-66DX2, a 3-4 year old CompuAdd HiRez VGA(Tseng ET4000) card,
a serial card(dual 16550s), and an SMC EtherEZ (ISA-based EtherCard
Plus Elite) net card. The EISA configuration utility for the AMI
BIOS - HiNT Corp. PEISA CHIPSET BIOS Rev. 2.10(1/17/94) does not
indicate any conflicts. In my research I found a note about a known
problem between ATI VGA cards and UltraStor having to do with ATI
improperly using some reserved addresses in the EISA slot-related
memory. I have tried the SCSI controller in a different slot and
still have the same problems. I also swapped out both the Seagate
drive and the UltraStor 24F with exact replacements and had the same
problems. I also tried using an APS Technologies active SCSI Sentry
with Digital Active Termination to try to ensure good termination on
the SCSI bus, but it made no difference.
Since UltraStor is apparently in bankruptcy and their phone numbers
are disconnected, I can't get help from them. I have been attempting
to get this configuration to run Linux error-free in my spare time
for several months now, and I would really appreciate any suggestions
you may have, including discarding the UltraStor if that indeed is my
only option. Sorry for the length, wanted to give as much info as I
could.
Thanks,