Running SLES11sp3 with OES11sp2 and GroupWise 2012sp2

My server started to to become unresponsive and GW stopped responding. So I ran a "rcgrpwise stop".
Groupwise shuts down and I then ran "shutdown -r now". The server starts to shutdown various processes and then does nothing.
So I wait a bit, then wait a bit longer (20 minutes) and still no reboot. I "ssh" into the box and try "init 6" and "init 0", but the server refuses to die.
I then hit the power button and the box restarts. All looks good but GW doesn't load....I can then see that the NSS volume has not loaded.
I then go into mild panic mode and need to run "ravsui" to verify the NSS pool, which takes 4 hours (It is not a peaceful 4 hours,
cos users start complaining about no mail). I then "activate the pool" and mount the drive with "nsscon"
Then all is good again .... until the next time.

So my question is how do I avoid this situation next time and be able to force the server to shutdown without breaking the links to the NSS volume?