More NFS client problems with 2.4.20

This is a follow-up to the previous post on NFS client problems with the RedHat 2.4.18 - 2.4.20 kernels.

After 3 weeks of no problems, the problem showed up again on a different cluster node. Four out of 12 nodes have now shown this problem, but it seems to affect nodes randomly. It always occurs during heavy NFS writes. A job running on the problem node stops, waiting for NFS read/writes. All of the other nodes continue to see the NFS server just fine. If I try to do an ls on the NFS mount point on the problem node, the shell locks up.

Checking nfsstat, retrans is not going up, however, getattr is increasing about 2 per sec. Also, the stuck processes always show a WCHAN of "end".

Any ideas what is going wrong?

Thanks,

Joshua Weage

=====

__________________________________
Do you Yahoo!?
The New Yahoo! Shopping - with improved product search
http://shopping.yahoo.com