SUMMARY scsi transport errors

From: Tom Trainor (tjtraino@to.mobil.com)
Date: Thu Nov 07 1996 - 08:56:20 CST


Managers:

Thanks to those who replied:

cedo@endeavour.exar.com
Cedo Vicente <cedo@exar.com>
Kevin.Sheehan@uniq.com.au (Kevin Sheehan {Consulting Poster Child})
dsf@interpath.com
Herbert Wengatz <hwe@uebemc.siemens.de>
dzambon@afit.af.mil (Dan A. Zambon)
root@wisdom.maf.nasa.gov (Mark Hargrave)

Here's the top two.... the first one worked....others suggested "check cables"
"check termination" etc. all of which I did twice after I got the
suggestions...

Number one answer begs another question:

I assume that in /etc/system I am changing or modifying kernel parameters
without having to recomplie kernal etc. What are the tree parameters specified
below AND (bigger question) where are ALL kernel parameters defined/explained.

By the way, this was a SunServer 1000 at Solaris 2.4 with an external 4.2 gig
Sun disk on a fast/wide scsi adapter (sorry I forgot to mention that).

-------------------------------------------------------------------------------
I had a similar problem with Solaris 2.5.0. I ended up added the following
command to the /etc/system file to get this fixed. I was told by Sun that 2.5.0
had alot of SCSI releated bugs and that I should upgrade to 2.5.1. After
upgrading to 2.5.1, the changes to the /etc/system were no longer required.

ADDED for Solaris 2.5.0

set sd:sd_max_throttle=10
set scsi_options=0x58
set scsi_reset_delay=10000

Cedo Vicente
-------------------------------------------------------------------------------
>From kevin@uniq.com.au Kevin Sheehan

You should probably include the other messages that come with this - if
you got something about "disconnected" or "tagged cmd" timing out,
then you may want to try turning off tagged queueing in /etc/system
by modifying scsi_options (flags are documented in
/usr/include/sys/scsi/conf/autoconf.h) but remember it will turn it
off for *everything*.

One thought really - more info. In general, you should tell us the machine,
the OS, what kind of disk &c. In this case, I'm taking a wild guess.

uname -a and the rest of the error log, as well as some info on the disk
failing would have helped a lot.
-------------------------------------------------------------------------------
>From dsf@interpath.com

I received the same message on my machine once. My problem was that I had
too many devices off one SCSI controller. I had:
      2 internal disks, 1 cdrom, 2 external disks, and 1 tape drive.
Normally this is okay...but the tape drive was inside a jukebox. The
jukebox device put me 1 over the limit. Moved jukebox and tape drive to
another controller and the problem disappeared.

Your problem might also be caused by the usual stuff--bad terminators,
cables, or loose connections. Good Luck.

Dan Freedman
-------------------------------------------------------------------------------

-- 

Tom Trainor Email: tjtraino@to.mobil.com Mobil Oil Corp. Phone: 703-846-3924 3225 Gallows rd / Rm 2B-308 Fax: 703-846-1460 Fairfax Va. 22037-0001



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:11:15 CDT