Written March 2009
by Cliff Feldwick
Working around the house, I used to call it “First, find the hammer.” You know, you go to do a simple task like hanging a picture, but the hammer is not on the bench in its usual place. That’s right; it was in the spare bedroom because you were fixing the shelves, or some such. But you stopped that because you needed some more wall anchors. So now you’re headed to Home Depot, but while you’re there you need more screws and the trashmen ran over the lid to the can so you need another of those. Oh, and since that’s near the post office you could mail that package to Henry and … and before you’re through it’s 2:30 in the afternoon and you’ve spent $87 and when someone asks what you’ve been doing all day, the answer is “Hanging a picture” except it still actually isn’t up.
This comes to mind because I’ve been watching as a client has tried to do an upgrade on their servers and has dropped down the same rabbit hole of “first, we have to load that upgrade”. This is now officially an educational tale, with a few thoughts of things you should consider before taking on a major upgrade or project.
I preface this by saying I am not a fan of changing things that work well. Often, the only really compelling reason to change from a working program or system is the usual creeping in of small-getting-larger problems that become distracting of your real mission. In their case, it was a combination of an update of the backup program, an absolutely necessary part of their daily routines, and the closing in on five year life of an important server. Dell will only take a hardware contract on servers for five years, and running without such reassurance was not an option.
Of course Microsoft, the new standard-bearer of planned obsolesce (replacing the car companies, who discovered that brands that actually made reliable cars had developed loyal followings who never wished to go back, thus crippling the old guard) has gone to whole new versions of their server software and Exchange, the program that runs internal e-mail, calendar management and such.
Anyway, the new Backup Exec wouldn’t run on the old platform until a service pack, the usual collection of bug fixes and security patches, was installed. Except it didn’t load correctly, necessitating calls to Microsoft and too many expensive hours of consultants making it work. And the new backup version needed tweaking.
That put the project behind. OK, moving forward, the new server hardware arrived and new versions of operating software were installed. So now comes the project of moving all the old e-mails to the new server. Much more complicated than originally anticipated. More hours planning and stomping bugs, and then it came time for the people who had Blackberrys. The present version of the software that transferred the messages from the office e-mail to the Blackberrys had to be upgraded to work with the new server. Suddenly half their messages from the Blackberrys were in Chinese. No really, the software was changing their messages to Chinese characters. Go back to Microsoft, delete the previous upgrade and re-load it, etc but no dice. Go to Blackberry and spend hours with their tech service people, keep slicing the onion down to smaller and smaller pieces until the root cause was found – a part of the upgrade that was supposed to automatically take care of this interface problem but didn’t. Manually go into the registry, the guts of Windows, and add a line. Eureka – it now works. What the final bill for all this high-octane consulting will total will probably knock down a mule. And the software from Verizon for the Blackberrys will have to be updated someday to tie up all the loose ends and lord knows what fun that will cause.
Oh yes, and the people who used to be able to get their office e-mail over the internet were locked out for days and unless you get a new security certificate from Network Solutions that costs $385 they’ll get a security alert that says they’re about to enter the danger zone. Nice.
So what can you learn from this? First, this is why consultants who give you an estimate on how many hours to complete a project will insist that it is only an estimate, and any attempts to limit the top end will likely fail. Experienced installers will have gone through this and know that the quickest way to kill any profit this quarter is to get trapped in a whirlpool of making something that should be “out of the box” work when it won’t.
Of course, if this becomes a normal part of working with them, you’ll need to re-evaluate. Just as you don’t want to own the first Volvo when a Ford mechanic decides he wants to learn about foreign cars, you want to establish that the consultants doing your work have real experience in the same work as your project. Get real names of past clients and ask if things went exactly as planned. If not, who paid for the experience?
And realize that the bigger the project, the more likely that Murphy’s Law will become a part of it. Brainstorm on what could happen, how to structure things to minimize havoc, including slicing things into smaller easier parts, and what to do if things take longer than anticipated. Either that, or learn Chinese.