[3226] in linux-scsi channel archive

home help back first fref pref prev next nref lref last post

Continuing troubles with Intel 4-CPU board with 2 788x's

daemon@ATHENA.MIT.EDU (Larry M. Augustin)
Fri Feb 6 02:57:22 1998

Date: 	Thu, 5 Feb 1998 17:18:26 -0800
From: "Larry M. Augustin" <lma@varesearch.com>
To: Jered J Floyd <jered@MIT.EDU>
Cc: linux-scsi@vger.rutgers.edu, linux-smp@vger.rutgers.edu
In-Reply-To: <s56yazpyie3.fsf@opus.MIT.EDU>


Yes.  See http://www.varesearch.com/products/vs4000.html.  I can tell
you that kernel compiles go really fast when you have 1GB of RAM and 4
processors.  It's kind of neat to compile the kernel and have Linux
cache the entire compilation in RAM :-)

I would help you if I could, but I'm not sure what to tell you to try.
We haven't used *exactly* that kernel, but we did use 2.0.31 on RH 4.2
with updated Adaptec drivers.  But that should all be in 2.0.32.  I
have not yet tested 2.0.33 on that system.

Also, we always use a RAID controller with the system and put the disk
array on the RAID controller.  We use the built-in Adaptec just for
peripherals like the tape drive.


Larry

Jered J Floyd writes:
 > 
 > Greetings,
 >   Last week I described troubles I was having with installing Linux
 > (specifically, RedHat 4.2) on a quad PentiumPro Intel motherboard that
 > has 2 on-board aic788x's.  At that point I suspected the problem was
 > due to there being no devices on the second controller, and in
 > response to my questions I got, "Install more drives" (not really an
 > option), and "Disable the second controller from the BIOS" (which
 > doesn't appear to be possible.)
 > 
 >   Since then, one of the drives has been moved to the second
 > controller, so each controller now has a drive.  The system runs fine
 > under Windows NT 4.0. (Well, as well as anything can run WinNT 4.0.)
 > 
 >   The symptoms as before are these: The devices on the first SCSI
 > controller (scsi0) are successfully detected.  Then the driver moves
 > on to the second controller, times out, faults and displays a stack
 > trace (that unfortunately scrolls away very quickly), spews several
 > "Idle task may not sleep" errors, and tries again.  It then gets stuck
 > in an endless loop of timing out and retrying on the second
 > controller.
 > 
 >    Has anyone been able to successfully install RedHat 4.2, or other
 > distribution that uses kernel 2.0.32, on this type of motherboard?
 > It's my understanding that it's a fairly common model.  Is there any
 > way to tell the SCSI driver to stop looking after the first
 > controller? Is there any way to make the driver stop retrying? (It
 > seems a bug to have it loop endlessly) Do I need a newer revision of
 > the aic788x driver? Any help would be appreciated.
 > 
 >    Thank you.
 > 
 > --Jered
 > jered@mit.edu

home help back first fref pref prev next nref lref last post