I'll be in Hanover on Thursday and Friday this week to teach Windows Server 2003, Active Directory and DNS to System Engineers of our company. Those trainings are pretty exciting - they usually don't know a lot about Windows Server 2003, and usually even less about Active Directory and DNS, but they have a very good understanding in different technologies (NT, Novel, SQL, ...) - whatever the job role requests at the customer they are assigned to.
I like to teach those classes - they are more forcing on me and less boring on me than if the audience has the same skill level. However I had to think about classes I had in the past. Class environments are usually totally different than production networks. I'm experienced in both, and I rarely get the same errors in a class than in production. E.g. usually in a classroom I have about six Active Directory domains with about the double number of DCs which we set up within one hour. There are effects doing this which you won't have in a production network where you have time, you ensure that a DC is working very well before you set up additional DCs. Further I've seen students integrating DNS into AD but doing it simultaneously on DCs of the same domain, and using different application partitions. Or one time I had a forest root failure a couple minutes before the class was supposed to start. Hardware failure. Unfortunately I had no other DCs in the root domain, but one sub domain (I was setting this up the day before, and did the first DCs first, then I let them replicate and went to print the handouts. After I got back the classroom was locked, so I couldn't add additional DCs on that evening). There was no time setting up the servers again, so I had to demote the sub domain, clean everything up, then using another Server as forest root and setting up the AD-Structure again. Somehow I managed to get everything done shortly and was able to start with the class. On the next day - same class - we heard a loud scratching and a weird sound. The new forest root server had a total hardware failure. This was pretty weird: OK, as in most classes we were using client hardware for the servers but two hardware failures on the forest root in two days. The students were pretty lucky - we had an additional DC in the root domain at that point and I demonstrated how to recover when the server holding all roles fails.
I'm looking forward to the class on Thursday. Hope to have some interesting chats and discussions with the students. And for some reasons classrooms are like geek-talk - you enjoy scenarios which you won't have in production or which you would never implement in production but you still want to know how to do it. Great!