Yesterday: One of my office domain controllers, ROOTDC01, failed. Not so much that things stopped working when it failed, but it left us open to serious downtime if it’s partner, ROOTDC02, failed before we had replaced the first one. I decided that it didn’t make sense to bring a replacement Windows 2000 domain controller in, only to proceed with our planned domain controller upgrade project in about 4 weeks. It only made extra work. This was the (sort of) perfect opportunity to bring in a shiny new already Windows 2003 DC into the organization. And it would also force me to finally “walk the walk” after quite a few months of “talk” (and testing!).
Earlier Today: This evening, a co-worker and I started on upgrading the schema in our organization to support this shiny new DC. This process, which happens on ROOTDC02 (the remaining DC), is relatively simple on paper and successful 99.9% of the time. But it could do major damage the other .01%. And since I didn’t have an 2nd DC to act as a backup, a screw-up could leave me doing a lot of disaster recovery. For many many hours.
All I really had to do was follow the step-by-step directions that I prepared for myself during the testing phases. And then, of course, second guess my directions. Wring my hands, close my eyes tightly and pace around while things were happening. And in some cases, when Microsoft documentation says “immediately” they really mean “give it 15 minutes to stew a bit.” This is when most of the pacing happens. And rapid refreshing of my replication monitor application.
Now: Everything seems to have gone nicely. No system errors that weren’t expected. No crashes, no blips. It’s only the year 2006 and I’ve finally gotten around to getting our systems up to 2003.
Monday: Bring in ROOTDC03, the new partner for ROOTDC02. We are still in a touchy spot over the weekend – but I think we’ll be fine. Once that new server is running, I can start upgrading the other DC to Windows 2003… might even finish the whole project before my deadline.