The Day the World Didn't End

Not sure if you are right or not about him in general. But he is spot on about Y2K.

I was a sysadmin for about 500 person site of an engineering company, Solaris, some Windows NT, and storage on some early NAS with a bespoke OS and hardware RAID card. We knew about it, were patched, tested, and ready to go by Dec _1998_. Yes, you can actually test date changes up front and not wait till midnight 2000! It was a about a weeks work, all told.

Then the corporate idiots started panicking around June of 1999. They hired a Y2K czar. He demanded reports! Plans! Contingency plans! He hired a dozen people for a department that normally had three. The resultant bureaucracy and meetings wasted 100% of now 16 people's time for 9 months.

And... Nothing was actually done, nor needed to be. Reports listing all the patches that had been applied a year earlier we demanded and made. Of course the answer "we're done, it's fine" was not conducive to career and empire building. So we kept turning in reports that said "we're done , go away". They kept getting translated into "we need to hire more managers because we can't get accurate status and no progress is being made".

And you know, the only thing I worried about was the Unknown-Unknowns, like a phone switches or alarm or power conditioners we didn't know about. Of course, the Y2K commission didn't know they existed. Thankfully the intersection of sets "things that even know the time" and "things that are hopelessly unmaintainable embedded devices" ended up to be empty.

The corporation sent out dire warnings in December to it's employees that their credit cards would stop working, they had to stock up on food, and quite possibly not show up to work for the month of January if things went badly. The engineers that worked there by then were having none of it and it was a joke.

My actual contingency plans involved dropping by to make sure everything was still alive on Sunday the 2nd, before everyone showed up to work on Monday. One seldom used license on a flexlm server was broken, but that was because it had an expiration date of 12-31-1999, not quite a Y2K thing. It was perpetual, and shouldn't have expired. But it correctly used a four digit date and knew what time it was. New license file was obtained in the next week. Nobody even noticed.

That was it. File storage, makefiles, incremental backups. All fine.

Did the bureaucracy that got created for this great effort disband? Nope. They continued the meetings for at least three months. They all got awards for doing such a great job. They stayed on to help guide and run the IT department indefinitely, due to such demonstrably good management. A short while later the commission pivoted to "Security" and "Anti-Virus" paranoia. Oddly that was the beginning of the decline in reliability and good IT at that site. Nothing ever really worked well again. It wasn't Y2K, it was having a 20 idiot managers instead of 3-4 guys that knew what they were doing. Within a couple of years we all quit.

Now, I've always assumed that somewhere there were actual hairy problems. Perhaps in a government or banking mainframe. It was said you could get rich writing COBOL. It didn't seem worth it.

But yes, it was a total non-event for workstation and PC based networks.