Just one more thing, an essay on troubleshooting

An keyboard and mouse

Over the weekend I watched — for the first time in my life — an episode of the long-running ABC and NBC police show Columbo, starring Peter Falk. I originally knew him from the film The Princess Bride (fun fact: I acted in a film directed by Buttercup herself, Robin Wright).

I’ve been hearing a lot about it recently because there’s a new streaming service in the USA called Peacock, owned by NBCUniversal, and I guess this is one of their flagship shows to convince people to sign up. Sadly, I am in Canada and the only way I can watch an episode is if it’s in the public domain.

It’s good TV. In the episode — directed by a young Steven Spielberg no less — Falk is still finding his character and so he lets the murderer know he’s on to him a little sooner than later on in the series, or so I’ve read, but it’s a remarkable performance from both him and the Big Bad, played by Jack Cassidy.

In a previous life before I joined the Microsoft Docs team as a technical writer, I was a consultant, dealing mostly with SQL Server on-premises, and specifically in performance tuning and disaster recovery best practices. Prior to that I had done a few turns in technical support, including telephone help desk and network administration. Over the years I became attuned to the systems with which I worked. I could tell when something seemed off. It’s like that alarm in your head when a colleague says, “I didn’t change anything.”

Frank Columbo is a scruffy raincoat-wearing police lieutenant working for the Los Angeles Police Department. He gets in people’s business, especially if they’re the killer in the episode. The episode format, for those of you who have never watched the show, is that the crime is shown in the first act. You know exactly who did it, and how they did it. It’s up to our scruffy lieutenant to figure out the motive and elicit a confession, because just like in real life, a conviction is much easier and cheaper to get in TV land if the murderer confesses.

In interviewing the prime suspect, Columbo keeps showing up to make the character feel uncomfortable. As we may have seen in our own lives, people who are aggrieved tend to show their hand, acting in a way that gives away their guilt. Perhaps it might border on harassment, but we should have some allowances for Peter Falk.

Columbo knows who the killer is early in the episode, and just needs to figure out the motive so that he can get the confession and roll the credits. His manner, his way of speaking, his behaviour, all comes across as bumbling, inept, disarming. He wants the killer to think he’s not very capable until he turns around and says his catchphrase “Just one more thing,” before dealing the final card that proves the case.

A system administrator who’s been on the job gets to know their environment. There’s that one server that needs to have the application pool restarted every 11 hours because the third-party library on Reporting Services has a memory leak. The CEO’s laptop always runs out of space twice a month because of Outlook, and that one server in the dev cabinet needs its %TEMP% folder cleaned out periodically.

As a technician in a computer store, you might get a laptop in that, as soon as you open a web browser, you know it’s been infected with malware. Maybe it’s the start page, maybe it’s the toolbars. Maybe it’s that weird icon on the desktop you’ve seen before. It’s one of a series of things, a trail of evidence, that points you towards the reason for the laptop being there.

In the SQL Server world, it’s the six ERRORLOG files. It’s the maintenance plans that were configured using the wizard. Maybe SQL Server Agent is still set to start manually. You just get a feeling. A vibe, as the cool kids say.

I’m writing this post today to ask you to trust that vibe. I started learning how to fix computers because when I had a problem with my computer as a kid, I didn’t know how to get it working and it was extremely frustrating. I went to a college to learn as much as I could in the shortest amount of time possible, and then came in first in that class. I then borrowed money and got an MCSE as well, still on the Windows NT 4 track. I realized that although the internals of computers were too small to see, I could understand them. I got their vibe. If a computer was going to crash, it would do so in front of me. My first paid job was on a technical help desk where I helped people install modems and set up Internet Explorer. I could see the screen in my mind.

I was a good network administrator, and all that help desk experience counted. I knew if a computer was infected with a virus or had file corruption. I could tell when a stick of RAM was bad. When I got into SQL Server performance tuning, I remembered all that stuff I learned before, and it all counted. Moving to Canada, the only job I could get was $17 an hour in a computer store where I removed viruses from slow laptops, and you can bet that being a former network administrator came in useful when I was able to share knowledge with my colleagues, and I learned from them too.

When something feels wrong with a system, trust the feeling. Like Columbo, you may not know the exact reason, but you start investigating. You keep going until the evidence, the motive, presents itself. Sometimes you’ll click on something you weren’t planning to. Maybe you find another problem that wasn’t even a consideration in the first place.

Share your accidental troubleshooting solutions in the comments.

Featured image by Hugo Barbosa.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: