Just chiming in here to provide some context. I had suggested stopping the service and verifying it was down prior to rebooting because I have seen (occasionally) in busier environments the service can take a little longer than expected to come down as it is trying to gracefully disconnect clients. That prevents corruption and the like but my theory was that it may not be terminating before Windows calls it quits and shuts down anyway. So far I've heard this has helped in Jay Dubb's environment, so this is something we can look at trying to reproduce in-house then get it over to development to see what we might be able to do about mitigating it in the future. For the moment, can you set up a script that stops the service, issues a wait, then proceeds to reboot the server normally?
Kyle Kerst
IT Coordinator
SmarterTools Inc.
www.smartertools.com