As can happen from time to time, there was not a single capture this morning when I went to look at the images. So I kicked off my usual trouble shooting routine.
First place to check is the captures page, so I can see when there was last a successful capture, which was the NOAA 15 pass starting at 21:15 on Friday night. So something happened after that.
Next place to look was the predictions page. The predictions for the day are calculated just after midnight each night and are then copied over to the web server. The predictions page still showed the page for Friday, so I knew that the problem started between about 21:35(ish) when the NOAA pass finished and a few minutes after midnight.
The next check is to see if I can SSH (Secure SHell) into the Raspberry Pi that controls the predictions and capture. This wasn’t possible from several servers on the network, so either the Pi was down or it wasn’t able to use the network.
I’ve got code in place on the Pi that if it can’t access the network or it can detect that other servers on the network can’t detect it, then it will restart networking or reboot. So clearly something had stopped this mechanism working.
So the next step is to go look at the Pi itself. It looked fine, but the case was cold to touch and slightly damp, which clearly shows that it wasn’t working for a while. A quick power cycle, i.e. remove the power and put it back, didn’t work. This wasn’t looking good as it could mean a dead Raspberry Pi.
Then I checked the power to the other Raspberry Pi that I have with multiple weather sensors. This was working just fine.
So I checked the power supply and found it covered in snails. There must have been about 10 all over it, no doubt attracted by it being warmer than anything around. The plug prongs had eased out from the socket, which explained everything.
A satellite weather station had been taken out of action by garden snails!
I’ve cleaned everything up and it is back on line, ready to capture the evening passes.