panic: spin lock held too long (w/ backtrace)

[Scott Swanson]

Kris Kennaway wrote:
> On Thu, May 10, 2007 at 02:49:45PM -0600, Scott Swanson wrote:
>> Hello all,
>>
>> I have a couple dozen SuperMicro servers in production and have had
>> reoccurring crashing issues on a couple of them that are under higher
>> load.
>>
>> After managing to pry one out of production, I have been able to
>> re-create the problem fairly consistently just by looping a script that
>> tars up a large directory and writes the the file to a different
>> partition on the same raid array.
>>
>> Can anyone point me towards the next step to debug this issue?  I can
>> pursue any other debugging info that may be required.
>>
>> Regards;
>> Scott Swanson
>> Inetz Media
>>
>>
>>> uname -a
>> FreeBSD pyrite.inetz.com 6.2-RELEASE-p4 FreeBSD 6.2-RELEASE-p4 #3: Tue
>> May  8 13:08:22 MDT 2007
>> root at biotite.inetz.com:/usr/obj/usr/src/sys/PYRITE  i386
>>
>>
>> root at pyrite /usr/obj/usr/src/sys/PYRITE > kgdb kernel.debug
>> /var/crash/vmcore.0
>> [GDB will not be able to debug user-mode threads:
>> /usr/lib/libthread_db.so: Undefined symbol "ps_pglobal_lookup"]
>> GNU gdb 6.1.1 [FreeBSD]
>>
>> Unread portion of the kernel message buffer:
>> spin lock smp rendezvous held by 0xc95a5900 for > 5 seconds
> 
> What is thread 0xc95a5900 doing?
> 
> Kris
> 

How can I determine the state of that thread?  Do I need to be using DDB
to glean that info, or can I use kgdb?

Can't seem to track this down in the Developers Handbook...

Regards;
Scott