Last night, I was patching a server for a new client. The server was built with SBS 2008, but had not been properly patched at all and as a result had given them no end of problems. I began the installation of Windows Server 2008 Service Pack 2 (SP2) onto the SBS 2008 server. Up until this point, the server had performed quite well. Naturally, I was doing this after hours, and given this is a client that works until 10:30pm, I was not able to start until after that. I was doing this remotely as the client is a 2 hour drive from home.
Anyway – the installation of SP2 normally takes around 45 minutes including the required reboots. The first 30 minutes or so of the installation went fine. After the reboot, the screen displayed “Installing service pack: Stage 3 of 3 – 65% complete" and it sat there. It sat there for another 30 minutes. By this time it was getting close to midnight and I was getting somewhat concerned as to why it was doing this. I started to review my fallback plans in case I needed to use them. Did I have a good backup? Of course – I did one just before installing the Service Pack using ShadowProtect. Did I disable the Antivirus software before installing SP2? Yes – given the scope of files being changed, I always aim to do this. I then started to look at the what if’s for the next morning. I figured that if there were issues, that I’d need to be onsite to resolve them – that would put a major dent in the day, but if things went wrong, what else could you do?
It was at this point that I stopped. I stood back from the problem and did some digging. I found reference on the Internet to someone who had experienced a similar issue. His server also stalled at 65% complete, and it was only while he was looking for his SBS 2008 DVD’s that it moved to 70%. I reasoned that given the server had given me no problems at all until this point, that it was likely that it was just slow for some reason. By this time however, it was well after midnight and the server had been at 65% for well over an hour. I decided to do the best thing I could do… and that was to leave it alone and go to bed. I figured I’d get up in the morning and see what the situation was and go from there. I knew I had backups. I knew I could restore the server if need be. I knew I could get the customer running in a few hours and there was no point in losing more sleep over it right now. So I went to bed. As it happened, the cat decided to wake me around 4am (bloody cat…grrr…) so I took advantaged of it to check on the server. Sure enough it had finished without an error.
The morale to the story is don’t rush things. Don’t go second guessing yourself. If you’ve done the preparation first with backups and the like, then you are well prepared for something to go wrong. Had I interrupted the system when it was stuck at 65%, it might well have turned out to be a recovery situation.