Adaptec 19160 Problems

Adaptec 19160 Problems

[Jason Perlow]

Greetings (and Linus and Alan if youre listening, thanks):

I'm having a few bizarre problems with an adaptec 19160 scsi controller and 
several Linux distributions inlcluding Redhat 6.2 and 7.0, and I was 
wondering if anyone encountered anything similar and might be able to help.

The machine is an Athlon 1.2Ghz with 512MB PC133 RAM, Microstar VIA KT-133 
Pro2 motherboard, with twin Adaptec 19160 32-bit SCSI untra 3 controllers.

Both scsi chains are properly terminated with U-160 rated cable, both drives 
are recognized by adaptec's bios and utilities. Irregardless of which brand 
of Ultra160 10K RPM hard disk I use (I have several set up in removable bays 
as bootable devices), during the setup routine and boot sequence of every 
distro I have tried with 2.2 kernels, the drives initialize but the bus 
appears to constantly reset and the drives timeout and re-try to initialize 
again and again and again according to console error messages when the 
Adaptec 7892 module attempts to load, and I cant progress to the 
partitioning stages. The only way to get around this is to use an IDE boot 
device and use the disk as secondary storage.

This appears to occur with Redhat 6.2, Redhat 7, Mandrake 7.2 and Suse 7.0 . 
  I know the hardware configuration is ok because I am using one of the 
removable drives with Windows 2000 and Windows ME just fine.

I have also used these same drives on Linux with other machines that have 
similar adaptec U160 chipsets (SGI 330, IBM Intellistation) and with the 
29160 controller (64 bit PCI version of 7892 chipset) without these 
problems, so this is puzzling the hell out of me. I have also tried ripping 
these controllers out of this new athlon box and tried them as bootable 
devices in a Pentium III machine, same problems.

Anyone know who is responsible for the adaptec module? I'd love to be able 
to kick this in the bud or find out if it was resolved in 2.4.0

thanks in advance!


Jason Perlow
President, Argonaut Systems Corp
http://www.argonautsystems.com
perlow@hotmail.com

Media Affiliations:

Contributing Editor, Sm@rt Partner     www.smartpartnermag.com
Sr. Technology Editor, Linux Magazine  www.linux-mag.com
Contributing Editor, Maximum Linux     www.maximumlinux.com

138 Vista Drive, Cedar Knolls NJ 07927
(973)451-0215 (646)349-1324 eFax

_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: Adaptec 19160 Problems

[I Lee Hetherington]

I had the exact same problem.  I ended up getting around it by
installing a Linux image from another machine.  2.2.18 works fine on the
machine, but Red Hat 6.2's install would not reliably get past the
infinite reset stage.

So, there is hope that once you get something new enough on the machine
you'll be in business.  You ought to be able to copy your working
IDE disk image to the SCSI disk, and with a new enough kernel boot the
SCSI disk.

--Lee Hetherington


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: Adaptec 19160 Problems

[Jason Perlow]

So you are saying this was fixed in 2.2.18? Which distro uses that by 
default now?

We need to get the distros to come up with boot floppy images for this then 
because 19160 is a very popular host adapter. Its not like its weirdo 
hardware. Waiting for an updated distro is a real pain in the ass if this 
can be fixed easily by just updating the boot disk images.

Jason


From: I Lee Hetherington <ilh@sls.lcs.mit.edu>
To: Jason Perlow <perlow@hotmail.com>
Subject: Re: Adaptec 19160 Problems
Date: Mon, 08 Jan 2001 14:45:26 -0500

Well, 2.2.18 is newer than what Red Hat 6.2 installer uses.

You could use Ghost or Drive Image to copy things over, or if you can
mount both simultaneously (booted from IDE), you can just make your SCSI
partition and cp -a everything over.  That is, everything except for
/lost+found, /proc, and any other mount points.

I agree this is ugly, but what can you do when the Red Hat (or
otherwise) install kernel doesn't really support your hardware?
I suppose it is possible to update the install kernel or specific
modules, and I think I once succeeded in doing this, but it is even
uglier!

--Lee


_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: Adaptec 19160 Problems

[Jason Perlow]

Ouch, thats an ugly solution.

But why would it be the installer routine as opposed to some wackyness in 
the adaptec module? The kernel used in the installer routines for most of 
these distros is the same kernel used to boot the installed OS, right?

How would you go about copying the IDE disk image anyways? With ghost?


From: I Lee Hetherington <ilh@sls.lcs.mit.edu>
To: Jason Perlow <perlow@hotmail.com>
CC: linux-kernel@vger.kernel.org
Subject: Re: Adaptec 19160 Problems
Date: Mon, 08 Jan 2001 13:39:37 -0500

I had the exact same problem.  I ended up getting around it by
installing a Linux image from another machine.  2.2.18 works fine on the
machine, but Red Hat 6.2's install would not reliably get past the
infinite reset stage.

So, there is hope that once you get something new enough on the machine
you'll be in business.  You ought to be able to copy your working
IDE disk image to the SCSI disk, and with a new enough kernel boot the
SCSI disk.

--Lee Hetherington



_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/