Active Pieces Maintenance in the Airport

I’m on my way to a tech conference this week in Las Vegas. My flight was delayed by five hours (sigh), so I had a little time to kill. Luckily I had my laptop and decided to fix up my rss2social workflow on Active Pieces. I got annoyed testing the workflow because every time I ran it, it would spam my accounts with posts that had already been published.

I decided to try Active Pieces’ table functionality: I initially loaded a CSV of every post I made on the site and had the workflow check the table first. If an entry wasn’t there, it would post and then add the URL for next time. I grabbed the site’s RSS file to get all the URLs and asked ChatGPT to convert it into a CSV I could import.

Active Pieces made the logic relatively simple, and I was able to update the workflow in about 10 minutes while sitting in a quiet spot at the airport. Woohoo!

Next up was restarting Active Pieces. Something is definitely wrong with my instance, because new posts don’t trigger its RSS trigger unless I log in and test. I also have a five-minute “health check” job that pings healthchecks.io every five minutes; it runs fine for about 12 hours and then stops.

To fix that, I needed to SSH into the Pi where Active Pieces runs, update it (there was an update I hoped would fix the stoppages), and restart. Except I couldn’t SSH into the Pi (over Tailscale) from the airport Wi‑Fi. Tailscale itself worked fine— I could access the Active Pieces web UI over it— so maybe the airport blocked port 22. Luckily I had a workaround: I could connect to my Mac Mini over Tailscale, open a terminal there, and SSH into the Pi to do the update and restart.

This is the first post since doing all that, so it’s partly a test to see if it will pick up the job and auto-post. Fingers crossed!