I was testing GroupWise on an OES2 SP1 Linux cluster with the cluster
nodes running as VMs under xen on SLES11. SBD and GroupWise resource
are located on an iscsi device.

These are fully patched SLES 10 SP2 with OES2 SP1 and GroupWise 8
virtual servers.

When I migrate or offline the GroupWise resource (an MTA and a Post
Office), ndsd on the hosting server just stops.

I've done a workaround by adding an "ndsd start" command at the end of
the unload script. This works fine.

Anyone have any idea what's killing ndsd?