I’m having the EXACT same issue, and it is annoying AF.
I recently moved from a RPi4 to a mini-PC as I wanted to get access to 16GB of ram. It was a rough migration, where my channels all got force closed, but no funds lost at least.
But on this device, I only have knots, lightning, tail scale, Thunderhub, and watchdog running, I don’t even have Electr, so bare minimum to see if it works fine.
I also have my older RPi4 still running, but not lightning app running.
Every other day, sometimes within a day, it just dies. The logs for Knots just show it stop logging.
I have to plug it out and turn it back on.
Once I got the following in my knots logs
2025-05-20T23:50:24Z Saw new cmpctblock header hash=0000000000000000000064b76273428397bd07367fc77f2499b4b6a0d723e110 peer=19
2025-05-20T23:50:25Z UpdateTip: new best=0000000000000000000064b76273428397bd07367fc77f2499b4b6a0d723e110 height=897620 vers>
2025-05-20T23:50:25Z runCommand error: system(curl -s -m 5 http://datum_datum_1:21000/NOTIFY) returned 1536
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@>
Never seen such a thing before. It’s almost like some final statement before dying. Whats most annoying is that my previous RPi4 just works without any issues. So the more up to date PC dies, and a 7+ year old Raspberry Pi 4 just works.
I’m thinking of moving my lightning account back to the old node until I can figure out whats going on, but any heads up as to whats going on, or at least what to look for to fix it? The logs in the User Interface seem not to go back beyond the reboot, so I can’t see whats going on before it crashes. I can look at the knots logs because I know where they are, but I don’t know where the logs are for the system itself.