Hopefully my post is in the right place and can engage your attention since it's really an urgent project, has anyone encountered similar issue before or got any idea what could be the problem? Feel free to let me know if any information valuable is missing. Thanks an awful lot.

My detailed information is as follows:

1.I created a NCS(Novell Cluster Services) 1.8.8 cluster using OES2 SP3 with SLES 10 SP3 x64 on two HP DL380 nodes
2.Ten logical volumes were mapped from a shared disk system to this cluster and 10 NSS volumes were created as per the policy that one separate LUN per pool and each NSS volume per pool
3.NFS load script(exit_on_error exportfs -o rw,sync,no_root_squash,fsid=101 *:/media/nss/VOLUME_A) and unload script(ignore_error exportfs -u *:/media/nss/VOLUME_A) is added to each NSS pool/volume to export NFS file system
4.These ten NFS file systems were mounted on a host and IO were run against these file systems


The host IO reports NFS stale error immediately after failover is performed on the cluster(reboot the master node)

I have done the following experiments but no luck:

1.change the remote_request_timeout parameter in /etc/evms.conf from default 12 to 1000
2.change the timeout value under Clusters->Cluster Manager->locate and select the name of the cluster->choose the pool name->Scripts->load script and unload script from 6 minutes to 10 minutes.