In the future do NOT make any changes to a production application server without:
1. Making an imager based backup so you can restore quickly in case of a SNAFU. This can be done with Symantec Ghost or numerous other imaging programs. If you have your server's system drives remapped, i.e. to M & N, then you'll need to use cmd line switches (-s -p -f -d, you can check these in the Ghost help menu)with ghost to preserve these settings, otherwise when restoring the image it will rever back to drive C and nothing will work.
2. Testing on a lab computer. You don't need an expensive server class machine to do testing. This can be done on an older PIII PC that you have lying around, or that you can get from ebay for $25.
3. In a large environment you'll do a pilot, i.e. update one server, watch it for a week before updating the rest of the farm.
I'm not picking on you, as we were all new at one time, just giving you advice for the future to avoid emergencies. If you have a fallback plan these problems are easily resolved by restoring the previous configuration and going back to your test machine to figure out the problem, while not interrupting business.
Patrick Rouse
Microsoft MVP - Terminal Server