kernel enhancement for stale NFS file handle?

Hi,

I am trying to configure Veritas Cluster server on RHAS2.1 (2 node configuration). Every thing works fine till I configure NFS service group. Once the service is up I am able to mount exported FS on my client machine. But once I switch over the service from one node to another, I get *stale NFS file handle* error for any open file (eg while doing dd if=/dev/zero of=<mntd_FS>/<file> I get stale handle for <file>) but I am able to do any new operation like ls, cd or dd without any need for unmounting/mounting the FS.
This requires some kernel enhancement and is implemented in RHAS2.1 (kernel 2.4.9-e.3) but still I am facing this problem.
I had same trouble with RED HAT CLUSTER MANAGER but somehow it got resolved. [ and I am not sure how :-( ]
Any idea how this can be resolved??

Regards
Chandan