[olug] Machine Locking up, update

[Miller, Scott L (Omaha Networks)]

Better link to the problem description, and includes links to patches that seem to fix things:

http://atlas.et.tudelft.nl/verwei90/nforce2/index.html

-Scott
-----Original Message-----
From: olug-bounces at olug.org [mailto:olug-bounces at olug.org]On Behalf Of
Miller, Scott L (Omaha Networks)
Sent: Tuesday, August 10, 2004 10:17 AM
To: Omaha Linux User Group
Subject: RE: [olug] Machine Locking up, update


Hi all,

Thanks for the suggestions.  Interestingly, Trent's comment about
"incompatible controller drivers for the mobo" is probably on the right
track.  For those interested, if you google for the following:
   linux nforce IDE drivers problems "lock up"
and click on the first link which should match something close to this:
   "Linux Kernel: Re: NForce2 pseudoscience stability testing"
There's a long thread I've been wading through that seems to be on the
right track to solving the issue.  I've not yet read the entire thread
so don't know if it's been solved yet.

As for your suggestions:
Jon, pings die when it locks up, and thanks for the 'X' comment, I'd 
not realized I hadn't tested without running 'X' yet.  It does take 
longer to freeze without 'X' running, but it does still freeze.

BTW, the way I'm getting the machine to lock up nearly on command is
to run 'fsck.ext3 -vtcc /dev/hda1' (the trashed root partition) or 
against /dev/hdc2 (my test partition).  It actually ran for 10 minutes
before crashing the first time running without 'X', so I began to get
worried that 'X' (->memory) did have something to do with it. It had
been crashing within 2 minutes of starting those tests lately.

It my last test for the night, I booted into knoppix run level 2,
started the fsck.ext3, then did a 'cat /proc/kmsg'. No messages 
were sent when it froze.  I don't remember seeing a dmesg, but didn't
have your email available when I got home :\

Eric, I plan on moving everything to one cable, then swaping cables
tonight.

With any luck, I'll get through the thread mentioned above and it will
give me something to try...

-Scott


-----Original Message-----
From: olug-bounces at olug.org [mailto:olug-bounces at olug.org]On Behalf Of
Eric Lusk
Sent: Monday, August 09, 2004 3:00 PM
To: Omaha Linux User Group
Subject: Re: [olug] Machine Locking up, need hardware guru advice


Aside from making sure it isn't the ribbon cable, I
would install just one hard drive and one cdrom on the
same controller.  See if it will run ok on the primary
alone, nothing on secondary, and vise versa.
Have seen a situation where the guy ended up getting a
PCI UDMA controller to keep his motherboard
functioning....


--- "Nathan D. Rotschafer"
<nrotschafer at geniussystems.net> wrote:

> May also consider compiling in serial control for  the box.  When I was
> having serious issues with one of my machines I did that and then hooked it
> up to the serial port of the other machine which lets you capture all
> messages and also shut it down correctly usually.
> 
> Just a thought,
> Nate
 
 
--- On 8/9/04 1:30 PM, "Jon H. Larsen" <relayer at levania.org> wrote:
> 
> Not sure how much this will help, but:
> 
> Put a bare-minimum linux distro on the system.  Make sure ssh is running
> and that your firewall will let connections in.
> 
> Try some copy operations, or play an mp3.
> 
> When you run into a lock, try to ping or ssh into the box and check dmesg
> or /var/log/messages for signs of problems.  When X locks up on me, I
> typically ssh in from remote and shut the machine down properly to avoid
> a possible Massive Filesystem Corruption (TM).
> 


-----Original Message-----
From: olug-bounces at olug.org [mailto:olug-bounces at olug.org]On Behalf Of
Trent Melcher
Sent: Monday, August 09, 2004 1:08 PM
To: 'Omaha Linux User Group'
Subject: RE: [olug] Machine Locking up, need hardware guru advice

> When you tested the secondary drive did you have this on the same ribbon
> cable as the Primary drive????   I have seen a bad ribbon cable exhibit
> this type of behavour, also incompatible controller drivers for the mobo
> can cause this....

> Trent

> Trent Melcher
> Network/System Administrator
> Startouch International LTD.
 
	
_______________________________________________
OLUG mailing list
OLUG at olug.org
http://lists.olug.org/mailman/listinfo/olug