One of the big hot topics around VMworld this year is Lab Manager 3. There is no doubt in my mind that this is a great tailored solution for any QA and Development environment. One thing that came to the forefront in many of the discussions around Lab Manager was the need for some up-front thinking/planning.
Sprawl is an issue with “virtual” machines anyhow. There is an overriding perceptions, particularly among users, that “virtual” mean costless. Of course we know better. But, Lab Manager can accelerate sprawl and proliferation of machines if it isn’t carefully planned and rolled out. In addition, it is really hard to go back and put this planning into place after the fact.
The first thing to consider is storage. Although Lab Manager does use the linked clones, it still has some pretty large space requirements. It’s also important to consider I/O performance. Big and cheap is probably not the way to go here. In terms of backup and performance, it also makes sense to carefully plan your data stores. You may also want to consider using extents to grow them as needed to accomodate linked machines which must exist together.
You also have to think about how your users are going to interact with the system. What level of ’self-service’ are you going to provide. You should also think about things like naming conventions, fencing requirements, and retention periods. Providing adequate training for everyone who will interact with the system is key. It should be part of the standard set of new employee orientation just like your other LOB systems.
In short, Lab Manger is great. But, it could faciliate a quick loss-of-control in your environment if you don’t take the time to do the neccessary up-front thinking. I look at Lab Manger, not just as a technology change, but as a culture change in your environemnt. You will get out of it what you put into it up front.
