|
|
|
remote control anxiety
Friday, March 22 2024
location: rural Hurley Township, Ulster County, NY
This morning I drove out to the Sherwin Williams on Albany (or is it Ulster?) Avenue to have them color-match the sage-green paint in our living room, which predates our keeping records of such colors. The guy there did a pretty good match and mixed me a quart of paint for use in painting exposed patches and new structures related to my ongoing lighting and fan installation work. I then went to Home Depot to get some electrical bits and pieces for finishing up some of the details of the fan installation.
Back at the house, I saw from the cabin data that the battery was at 100%, meaning any extra electicity being gathered was being wasted. So it seemed like a good idea to run the boiler for awhile, as temperatures in the cabin had fallen to about 49 degrees. So I initiated the heating using my system. It took nearly an hour to start, suggesting there was some sort of problem with the way the NodeMCU was recieving data. So I created another NodeMCU running the same program and ran it in the laboratory. The internet environment here has a much faster connection, so the program polled the server much more frequently than the server is polled from the cabin with an identical configuration. But I was seeing it almost never seemed to receive JSON data. At the cabin, the frequency of receiving JSON data had been much higher, something like half the time, though this rate did seem to decrease the longer the NodeMCU ran. Perhaps after a few days it happened only once every hour, which certainly isn't ideal. What could the problem be? It looked like I had more work to do.
Then, some hours later, I wanted to turn the heat off. But again, it refused to do so for well over an hour. The sun was going down and now I the boiler was eating into the power in the battery, power that needed to be there if, as the forecast predicted, snow came and covered up the panels again. At this point I was getting desperate and trying various tricks to make the NodeMCU reboot itself, much the way the Jet Propulsion Laboratory does with an unresponsive space probe. One such technique was to simply turn off the Apache dæmon on the server, thereby rendering it unreachable. After a certain number of connection failed attempts over a five minute period, there's supposed to be a reboot. But judging from the data being logged in the Apache access logs (which includes a log of failed connection attemps that clears after a reboot) that didn't seem to be happening to the NodeMCU doing the remote control. All of this was a valuable experience, because it opened my eyes to new features to include in the code and the importance of reliably getting data from the server. But it was also causing me a great deal of anxiety. Because if the boiler couldn't be turned off, it would burn through the battery well and render the cabin powerless in less than a day.
Meanwhile Gretchen had gone to another Woodstock Bookfest event at the Bearsville Theatre. (She'd originally planned for me to come as well, but after last night's experience, I knew there was no way I was going to sit in a crowd of people facing a stage for two evenings in a row.) I should've probably gone to the event tonight instead of the story slam, because it was to be a significantly more interesting presentation by Masha Gessen, one of Vladamir Putin's many enemies-in-exile. (We'd joked yesterday about not standing too close to them in case one of Putin's goons jabs them with the radioactive tip of an umbrella. I used those pronouns because those are the ones Gessen prefers.)
To calm my nerves and sooth my remote control anxiety, I took a bath. While there, I mostly made myself not check the data from the cabin. But after the bath when I finally did, I saw in the data that the cabin boiler had finally responded to my shutoff command.
For linking purposes this article's URL is:
http://asecular.com/blog.php?240322 feedback
previous | next |
