-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Your problem sounds very simular to what I have experienced.
I found back in Sep
http://marc.theaimsgroup.com/?l=linux-xfs&m=100008724000734&w=2
that under some tests the machine would hang or die. This would happen when
there were many processes moving data around the XFS volume.
I originally found this when I had heaps of NFS and SMB clients uploading
data at the same time. I was able to narrow the problem down to just a
kernel-xfs issue and could simulate the behaviour using a simple shell script
to start many background cp procceses. What would happen is kupdated would
seem to go into deadlock and the machine would stop responding. The machine
could still be pinged on the network but console or ssh would not be
responsive.
Then reciently I was building a few new file servers so I decided to get the
latest XFS kernel (2.4.16-xfs) and try again. A simular problem occured.
After disscussions on this list
http://marc.theaimsgroup.com/?l=linux-xfs&m=100855332323037&w=2
I tried yesterday's CVS kernel and everything is now running sweet.
I have yet to rerun all my tests but at the moment it seems to pass the tests
that use to kill it.
When you looked at the processes in the Alt-SysRq output was kupdated one of
the processes running? If it was I expect you have just hit the same problem
I had.
- --
Adrian Head
(Public Key available on request.)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org
iD8DBQE8IJ5Q8ZJI8OvSkAcRAmQQAKCRbLYg2u3lZBfuLdHx1LoI1LwP9ACeJNwy
eQfNH+1HJFAYMuSY+DZ2AOA=
=XDjc
-----END PGP SIGNATURE-----
|