Dave Kleikamp wrote:
> On Thu, 2007-05-17 at 09:37 -0500, Roger Heflin wrote:
>> Dave Kleikamp wrote:
>>
>>> I don't have an answer to an ext3 deadlock, but this looks like a jfs
>>> problem that was recently fixed in linux-2.6.22-rc1.  I had intended to
>>> send it to the stable kernel after it was picked up in mainline, but
>>> hadn't gotten to it yet.
>>>
>>> The patch is here:
>>> http://git.kernel.org/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=05ec9e26be1f668ccba4ca54d9a4966c6208c611
>>>
>> Dave,
>>
>> That appears to have fixed the JFS hangup.
>>
>> MTBF before was about 1 hour, under the same test I am over 20 hours
>> and things appear to still be holding together.
> 
> Great.  The patch is queued for the 2.6.21 stable tree now.
> 
> Thanks,
> Shaggy

Dave,

Apparently there appears to be another different similar lockup,
The MTBF has risen from 1-2 hours without that patch to >100 hours,
so I am fairly sure the patch did correct the original lockup, or
at the very least make it a lot less likely.

I hit the machine across NFS for 5 days before it deadlocked, before
the patch I could only get an hour or two (2-4 different tries).

Given that pdflush is "D" it does not appear to be an NFS issue.

Included is the sysrq-t.

This is with 2.6.21.1 + the JFSIO patch.