This issue appears to be resolved as of kernel 2.6.18-308.4.1.el5
We have a few Redhat Enterprise Linux Servers (5.8 64-bit) that perform some nightly backups by copying a large amount of data via NFS to another RHEL server.
Since upgrading these servers to RHEL 5.8 we’ve run into a problem where shortly after the NFS copy operation the client servers freeze. We are unable to ping, SSH or even use the console to access the server after the freeze. We are forced to power cycle them.
This issue occurred only after upgrading to RHEL 5.8 and while using the new kernel that came with it (2.6.18-308.1.1.el5).
We found reverting the kernel on the clients to 2.6.18-274.18.1.el5 solved the problem.
A bug has been submitted here: https://bugzilla.redhat.com/show_bug.cgi?id=805660 and hopefully I’m correct in thinking it’s a kernel bug.
We left our NFS server running the new 2.6.18-308.1.1.el5 kernel and it’s been fine.