From tege@matematik.su.se Mon Mar  1 10:53:00 1999
Return-Path: <tege@matematik.su.se>
Received: from pavidus.matematik.su.se (pavidus.matematik.su.se [130.237.198.6])
	by hub.freebsd.org (Postfix) with ESMTP id 0609D153D7
	for <FreeBSD-gnats-submit@freebsd.org>; Mon,  1 Mar 1999 10:52:15 -0800 (PST)
	(envelope-from tege@matematik.su.se)
Received: from tiger.matematik.su.se (tiger.matematik.su.se [130.237.198.18]) by pavidus.matematik.su.se (8.8.8/8.6.9) with ESMTP id TAA25166 for <FreeBSD-gnats-submit@freebsd.org>; Mon, 1 Mar 1999 19:51:57 +0100
Received: from tiger.matematik.su.se (localhost [127.0.0.1])
	by tiger.matematik.su.se (8.9.1a/8.9.1) with ESMTP id TAA28708
	for <FreeBSD-gnats-submit@freebsd.org>; Mon, 1 Mar 1999 19:51:58 +0100
Message-Id: <199903011851.TAA28708@tiger.matematik.su.se>
Date: Mon, 01 Mar 1999 19:51:57 +0100
From: Torbjorn Granlund <tege@matematik.su.se>
To: FreeBSD-gnats-submit@freebsd.org
Subject: System freezes during certain SCSI activity

>Number:         10332
>Category:       kern
>Synopsis:       System freezes during certain SCSI activity
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    gibbs
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Mon Mar  1 11:00:01 PST 1999
>Closed-Date:    Sat Feb 24 11:39:57 PST 2001
>Last-Modified:  Sat Feb 24 11:40:26 PST 2001
>Originator:     Torbjorn Granlund
>Release:        FreeBSD 3.1-RELEASE i386
>Organization:
Swox AB
>Environment:

motherboard:	ASUS P2B-S (PCB rev 1.02, BIOS rev 1008)
cpu:		Intel Pentium II 400MHz
memory:		256MB SDRAM w ECC (2x128MB Samsung GH)
disk:		9.1GB U2W Seagate Cheetah (da0, SCSI id 0)
disk:		9.1GB U2W Seagate Cheetah (da1, SCSI id 1)
disk:		9.1GB U2W Seagate Cheetah (da2, SCSI id 2)
disk:		18.2GB U2W Seagate Barracuda (da3, SCSI id 3)
tape:		12GB/24GB DDS-1/2/3 FSCSI Sony SDT-9000 (SCSI id 6)
scsi:		AHA7890 on motherboard
video:		S3 Trio64
network:	Intel Etherexpress Pro 100B

>Description:

When accessing the tape drive using the `dump' command, the system hangs
within a few hours.  The console then has messages like this:

(da0:ahc0:0:0:0) SCB 0xd - times out in dataout phase, SEQADDR == 0x5d
(da0:ahc0:0:0:0) SCB 0x5d - times out in dataout phase, SEQADDR == 0x5d
(da0:ahc0:0:0:0) SCB 0x13 - times out in dataout phase, SEQADDR == 0x5d
  .
  .
  .

The system then responds to keyboard input but any disk activity seems to be
blocked.  We have waited several minutes but the system remains locked.
Occasionally, more messages like the ones above are printed.  The disk
mentioned in the messages are sometimes da1 or da2.

The SCSI buses are properly terminated.

We cannot seem to provoke the problem unless the tape drive is accessed.

(We tried to connect a 2nd SCSI controller [an ASUS SC200/Symbios] to a PCI
slot, but then the system hangs during boot.)

>How-To-Repeat:

Happens every time we try to run a full dump of our disks.  Sometimes it
takes several hours, sometimes it happens sooner.

>Fix:

>Release-Note:
>Audit-Trail:

From: "Kenneth D. Merry" <ken@plutotech.com>
To: tege@matematik.su.se (Torbjorn Granlund)
Cc: FreeBSD-gnats-submit@FreeBSD.ORG, gibbs@FreeBSD.ORG
Subject: Re: kern/10332: System freezes during certain SCSI activity
Date: Mon, 1 Mar 1999 17:15:37 -0700 (MST)

 Torbjorn Granlund wrote...
 > 
 > >Number:         10332
 > >Category:       kern
 > >Synopsis:       System freezes during certain SCSI activity
 > >Confidential:   no
 > >Severity:       critical
 > >Priority:       high
 > >Responsible:    freebsd-bugs
 > >State:          open
 > >Quarter:        
 > >Keywords:       
 > >Date-Required:
 > >Class:          sw-bug
 > >Submitter-Id:   current-users
 > >Arrival-Date:   Mon Mar  1 11:00:01 PST 1999
 > >Closed-Date:
 > >Last-Modified:
 > >Originator:     Torbjorn Granlund
 > >Release:        FreeBSD 3.1-RELEASE i386
 > >Organization:
 > Swox AB
 > >Environment:
 > 
 > motherboard:	ASUS P2B-S (PCB rev 1.02, BIOS rev 1008)
 > cpu:		Intel Pentium II 400MHz
 > memory:		256MB SDRAM w ECC (2x128MB Samsung GH)
 > disk:		9.1GB U2W Seagate Cheetah (da0, SCSI id 0)
 > disk:		9.1GB U2W Seagate Cheetah (da1, SCSI id 1)
 > disk:		9.1GB U2W Seagate Cheetah (da2, SCSI id 2)
 > disk:		18.2GB U2W Seagate Barracuda (da3, SCSI id 3)
 > tape:		12GB/24GB DDS-1/2/3 FSCSI Sony SDT-9000 (SCSI id 6)
 > scsi:		AHA7890 on motherboard
 > video:		S3 Trio64
 > network:	Intel Etherexpress Pro 100B
 > 
 > >Description:
 > 
 > When accessing the tape drive using the `dump' command, the system hangs
 > within a few hours.  The console then has messages like this:
 > 
 > (da0:ahc0:0:0:0) SCB 0xd - times out in dataout phase, SEQADDR == 0x5d
 > (da0:ahc0:0:0:0) SCB 0x5d - times out in dataout phase, SEQADDR == 0x5d
 > (da0:ahc0:0:0:0) SCB 0x13 - times out in dataout phase, SEQADDR == 0x5d
 >   .
 >   .
 >   .
 > 
 > The system then responds to keyboard input but any disk activity seems to be
 > blocked.  We have waited several minutes but the system remains locked.
 > Occasionally, more messages like the ones above are printed.  The disk
 > mentioned in the messages are sometimes da1 or da2.
 > 
 > The SCSI buses are properly terminated.
 > 
 > We cannot seem to provoke the problem unless the tape drive is accessed.
 > 
 > (We tried to connect a 2nd SCSI controller [an ASUS SC200/Symbios] to a PCI
 > slot, but then the system hangs during boot.)
 
 This sounds like a cabling problem.  The 'timed out in dataout' phase means
 that the SCSI bus was in the data out phase when the 60-second I/O timeout
 hit.  This generally means you have a cabling problem.  (because a signal
 is stuck somewhere and the data transfer didn't complete)
 
 You should check for bent pins on the cable, make sure the cables don't
 run too close to the power supply, and make sure you're using the right
 type of cable, especially for the LVD portion of the bus.  The LVD cable
 should be one of the "twisty" cables with the terminator block at the end.
 
 That is assuming, of course, you're running at LVD speeds.  No matter what
 speed the LVD drives run at, they'll need external termination, because
 they don't have onboard terminators.
 
 Ken
 -- 
 Kenneth Merry
 ken@plutotech.com
 

From: Torbjorn Granlund <tege@matematik.su.se>
To: "Kenneth D. Merry" <ken@plutotech.com>
Cc: FreeBSD-gnats-submit@FreeBSD.ORG, gibbs@FreeBSD.ORG
Subject: Re: kern/10332: System freezes during certain SCSI activity 
Date: Tue, 02 Mar 1999 01:40:24 +0100

 Ken,
 
 Thanks for the reply!
 
 Yes, the timeout would sound like a cabling problem to me too.  But the fact
 that I have run the system w/o any problems with the same LVD cabling for
 1/2 year tells me the cabling is fine.  I added the tape unit lately, and
 that on the SE bus.  That started the problems, but I am unwilling to remove
 the tape unit.
 
 And yes, I have proper LVD cabling for the 4 disks (a messy cable with lots
 of twisted pairs...).  The tape unit is connected with a plain old 50-pin
 cable.  I have terminators in place (on the cable for the LVD bus and in the
 tape unit for the SE bus).
 
 If this is some sort of cable problem, why does it take hours to make it
 happen?  And why cannot FreeBSD recover from the problem?
 
 I have had cable problems in the past (with plain Ultra devices), and then
 FreeBSD was always able to recover (within a minute).
 
 Torbjrn
 

From: Joerg-Henrik Wist <wist@informatik.tu-muenchen.de>
To: freebsd-gnats-submit@freebsd.org, tege@matematik.su.se
Cc:  
Subject: Re: kern/10332: System freezes during certain SCSI activity
Date: Thu, 10 Jun 1999 17:47:45 +0200

 Hi,
 I have a similar problem, which I thought I should mention although I
 can't give any exact details right now (as the box is at home and I
 am at uni at the moment)
 In short: System is a K5/166, 64MB Ram, NCR53C810 SCSI controller 
           - 2.1G Quantum Fireball (da0)
           - 4.5G IBM  (da1)
           - YAMAHA CD-RW 
           - SONY SDT7000 DD2-Tape (sa0)
           - FreeBSD 3.1 RELEASE (waiting for the 3.2 CDs), custom kernel
 
 The CD-RW and Tape are in an external SCSI-Case, the two harddisks in
 the tower. Everything is working fine, except when I try to access the
 tape. Writing with dump is ok, but reading with restore freezes the box
 and on the console I can read: nrc0 timeout ... (don't recall the exact
 message). Reboot is the only thing that helps. Note: I can't think of a
 cabling problem, as for example  burning CDs works just fine. Could it
 be the driver for the tape?
 
 If you need any further info, please mail me and I'll try my best.
 Thanks, Henrik
 
Responsible-Changed-From-To: freebsd-bugs->ken 
Responsible-Changed-By: msmith 
Responsible-Changed-When: Wed Jun 23 17:01:32 PDT 1999 
Responsible-Changed-Why:  
Ken fielded this one; I suspect it may be resolved already if it was a 7890 issue. 
Responsible-Changed-From-To: ken->gibbs 
Responsible-Changed-By: ken 
Responsible-Changed-When: Wed Jun 23 17:10:28 PDT 1999 
Responsible-Changed-Why:  
I think Justin would probably have a better chance of resolving this 
than I would. 
State-Changed-From-To: open->closed 
State-Changed-By: gibbs 
State-Changed-When: Sat Feb 24 11:39:57 PST 2001 
State-Changed-Why:  
I believe this has long since been fixed. 

http://www.freebsd.org/cgi/query-pr.cgi?pr=10332 
>Unformatted:
