lol... I'll reserve my comments, as sil pretty much summed up my thinking, and his reply was funnier than mine, anyway! 

I don't get offended by words on the contrary, I was pointing out something you may have overlooked so you could come around and clarify what it was you were trying to accomplish.

The problem with trying to mimic a production machine to test against is... You're STILL not testing the same machine. For example, you have two similar clocks, in fact, they're the same exact clocks. You take one apart because you don't care what happens to it at the end of the day, its not your primary clock. You do this because you believe that by doing so, you will understand perhaps the problem with the primary clock. You also want to remember how to put the pieces back together again.

You spend all this time on your backup clock. Learning its ins, outs, how the hands work, how to take it apart, how to put it back together. Guess what you've done? You've done nothing more than worked on a clock that means nothing at the end of the day. While theoretically the primary should work the same, there is a caveat called Murphy (http://en.wikipedia.org/wiki/Murphy%27s_law). Anything can skew the true results of what you would see in the primary that would be missed simply because you never did any testing on it. You assumed that by having the secondary mimic clock (the dummy webserver) that things would be the same. It will NOT be the case.

Furthermore, so what you took the clock apart and put it back together. What did you learn about it. Can you rebuild the clock on your own from scratch in the event you don't have a copy? There are a lot of things that you need to know, in and out on the machine that COUNTS not the replica. Which is why you were told via Hit_Monkey, you should seek a professional