September 2, 2005
** Listen to this
column on your computer, iPod or other audio player **
As if I needed it, this week I had an object lesson in
the important principle of “Don’t Wait!” I try always
to take this to heart and usually manage to operate under its precepts,
but a combination of events led to a series of yet other events that needn’t
have happened. Thank goodness, the damage to data, and self-esteem, was
minimal, but it teaches an important lesson that bears repeating.
Do the project now
This particular event has its roots in a new client I first spoke to a
few months ago. They had a small office running an aging Novell server
using Groupwise as their email. Together we developed a plan to replace
this server, upgrade their network and replace their email with an out-sourced
solution. The company was not ready to move on the project immediately
for a variety of reason, so I estimated the cost of the project and left
it up to them to contact me when they were ready.
This week, during the height of a heat wave here in Los Angeles, I received
an emergency call from the client. Their server had crashed. I knew that
the machine was in a location that had very bad ventilation, so I told
them to move it into the office and let it cool down for a while. Sure
enough, after a few hours we were able to reboot the server and go back
to work. This event brought home the need to move forward on the project
Do what you can
We arranged to have a new server delivered overnight and I quickly backed
up the existing files to a local desktop machine as a safeguard against
any future failures. I had not yet setup the new email addresses, though,
so I postponed exporting the email from Groupwise until I had a place
to put it. It is here that I broke my own rule. I assumed the server would
last for 1 more day and I would be able to move the email out in an organized
fashion. Of course, I was wrong. The server crashed again the next day
and we were unable to resurrect it a second time.
What went wrong?
There are several points of failure illustrated in this story, all of
which you will face at one time or another. First, clients are often reluctant
to replace aging systems, especially if they appear to be functioning
well enough for their needs. They seem to want to wring every bit of cost
out of the system before committing to a new one. In this case, while
the server had continued to function, both their UPS and tape backup unit
had failed months before. This is not an unusual situation, though. I
often find such things when I meet with a client for the first time. That
said, these earlier failures should have created more of a sense of urgency
to replace the system.
Secondly, when confronted with the initial failure, I should have exported
all the mail to a safe location, just as I did with the data files. Instead,
I was trying to hold on to an old plan instead of reacting to new issues
immediately. I had wanted to move through this project slowly, running
the new systems in parallel and therefore execute a smooth and elegant
transition. In trying to stick to that original plan, it opened a window
to data loss. This is a perfect example of what can happen when we are
not open to a rapidly changing environment.
Third, the existing equipment was located in a room with little to no
air-conditioning and surrounded by other equipment. This location should
have been questioned during the initial installation. It is the job of
all high-tech workers to help their clients avoid preventable problems.
Somewhere in the past, changes could have been that might have postponed
or prevented this failure. Be kind to your clients, and your fellow high-tech
peers, when installing or updating new systems. It is important to remember
that you might be reviewing or replacing their work in the future.
Hindsight is often the best teaching tool, if also the cruelest. I hope
you find my experience a useful reminder that there are times when projects
should be pushed through, regardless of whether the current system is
still functioning. As high-tech careerists, there are times when we must
impress a sense of urgency on our clients. Both for their benefit and