The more tech you allow into your life, the more time you’ll spend fixing tech.
Sigh. Yesterday I discovered that there’s a crucial component in this whiz-bang site management system that I had not considered. How did I discover it? Because it failed. So now progress on all other fronts has to take a back seat while I figure out how to either repair or replace it.
And the tech in question? A tiny Raspberry Pi.
◇◆◇◆◇◆◇◆◇◆◇◆◇◆◇◆◇◆◇
The crucial cog in my new dataflow is the part where my Obsidian notes get automatically synced to all my computers. I’ve got SyncThing set up on all my devices, and they all point to the Pi, which acts as the central server for important operational files in my world: my git repos, my Obsidian vaults, etc. But it stopped working on Monday, meaning I can no longer test changes to WebSmith, because I can’t feed new test pages to the data chain.
I can’t even post more articles to this site to tell you all what’s going on. (Obviously, if you’re reading this, I must have figured it out, but at the time I’m writing, I’ve still got no clue what the problem is.)
All I know is that a simple reboot wasn’t enough to bring it back, which means this is going to require a bit more investigation. But before I can do that, I first have to set up somewhere to actually work on it, because my new electronics studio is not yet a reality and all my tools are still in storage boxes.
Creating Elbow Room (Day Two)
Those boxes I mentioned are piled everywhere in my new space, but remember those workstations I was talking about a while back? I had already built two of the carcasses and had parts on hand for two more, just waiting to be assembled, but this morning I had a brainwave. Instead of assembling the third one as a workstation, I assembled it as two low wheeled carts. That allowed me to get all the boxes up onto rolling storage, making it much easier to move them around. I’ve now cleared 75% of the floor space, which feels like an enormous accomplishment. I can actually move around in there without first moving the boxes ahead of me to make a path.
Tonight I’ll clear space on the work table, set up my computer forensics station, and see if I can get a pulse from the Pi…
2 hours later… Oops. Slight roadblock: I can’t find a keyboard to plug into the Pi. I know I’ve got a few, but they’re in a box. :-) And finding that box among the other boxes just became a project for tomorrow.
All Clear
Whew! It took a frustrating amount of effort to get things sorted out, but once I did, I was relieved to discover that the Pi itself did not actually fail. It was the power supply, which had developed an intermittent undervolt problems that was causing spurious reboots. If it had failed outright, I would have spotted that immediately, but inconsistent voltage drops made it harder to diagnose.
But I can now report that the data chain has been restored. And as a bonus, this episode has also alerted me to the potential failure point. I should probably take steps to ensure that, if it fails again, I’ll be able to swap it out in relatively short order. But that’s a topic for another time.