Subject: Re: Update: MP AS4100 hanging
To: None <port-alpha@netbsd.org>
From: Jeff Workman <jworkman@pimpworks.org>
List: port-alpha
Date: 01/08/2003 23:08:05
Stoned koalas drooled eucalyptus spit in awe as Daniel Carosone exclaimed:

> On Wed, Jan 08, 2003 at 06:59:09PM +1100, Luke Mewburn wrote:
>
>> Koalas are marsupials, not mammals (which bears are).
>> There is no such thing as a "koala bear".

You guys are too technical sometimes. :-P

Actually, I'd been corrected on that some time ago, and thought I'd changed 
it. Guess I hadn't.

Anyhow, I've been trying to track down this problem.  I'm not a programmer 
and definitely haven't much knowledge of kernel internals, so I'm fairly 
limited in what I can do.  Further compounding the problem is the fact that 
I'm not getting any kind of kernel logs when the machine locks up, nor is 
the console dropping to the debugger.

At first, after upgrading to 1.6L cvs, I thought that the hangs were only 
happening during NFS I/O.  I was wrong.

I was able to get the machine to hang by doing the following:

FTPing a large file from the machine that was on a JBOD hanging off the 
ISP1020.

FTPing a large file from the machine that was on a RAID5 array hanging off 
the DAC960.

Making the RAID5 array available via NFS and copying the file to a remote 
machine.

Moving the file from the JBOD to the RAID 5 array.

It doesn't look like heavy network I/O by itself is causing any problems. 
As I type this, I've got 2 "ping -f's" banging on the machine from two 
other machines on the LAN.  They've been running about 15 minutes without 
any problems.  I will continue to bang on the machine for another hour or 
so to see if I can crash it.

-Jeff
--
Jeff Workman | jworkman@pimpworks.org | http://www.pimpworks.org