I still think drfresh is heading for trouble, especially on a production server that's also been used for testing, i.e. rpms installed/removed, configuration changed and changed again. I hope drfresh will keep the community apprised of how the project goes and prove me wrong with a step-by-step account of what he did and how he did it ... but frankly I am still skeptical. I suppose the point at which it would have been easier and less time-consuming to start fresh than to attempt an "upgrade" does vary from person to person; me, I value my time greatly and would prefer to do the job and put it behind me.
The law of unintended consequences can bite you when you least expect it ... my latest SME issue in that regard was what I thought would be a simple upgrade from 5.5 to 6.01-01, on a bone-stock server that had been running perfectly for months. (After the fact I found out I should've done the 5.5->5.6->6.01-01 route but too late, since I was overconfident and didn't copy the mirrored drives.)
After the upgrade, everything worked perfectly ... except ... no workstation on the LAN could get DNS resolution. Period. Ping by IP, fine. No name resolution ever, even when ISP's DNS server's explicitly added. After posting that question on the forums and reading about a couple of other people with the same issue, I chose to back up the files, reformat & install a fresh 6.01-01. Now I don't give that server a second thought ... it just runs.
In drfresh's scenario, if it works at all he'll have to watch it like a hawk for a month, at least, to make sure there's no subtle glitch that escaped his immediate attention. A shared printer problem, perhaps, or a directory permission that's gone awry, or a contrib that left some template modifications behind, or a backdoor into the system left behind by a partially-removed PHP-based contrib.
Or he can just hope that absolutely nothing goes wrong ... or that the users of this production server are the understanding and sympathetic type.
So I stand by my original post, in which my intention was to discourage a risky undertaking on a production server. But since it's going to be transferred and upgraded on new hardware, I'll say "go for it mate but don't reformat that old server just yet!"
Good luck! Keep the thread alive by posting your results, preferably in detail so others can learn from your feat.