I recently ran into an issue where my master control process was crashing. I haven’t yet figured out the root cause, but it runs fine for 12-24 hours before crashing. After crashing, I have been able to restart the process without issue. As a stop-gap measure, I decided to set up the process to automatically restart in the event of a crash…
To do this, I created a systemd service by creating the following file:
/lib/systemd/system/thermo.service
[Unit]
Description=The thermoPi master control service
After=multi-user.target
[Service]
ExecStart=/usr/bin/python /home/pi/thermoPiDev/thermo/control/master.py
Restart=always
RestartSec=1
[Install]
WantedBy=multi-user.target
Alias=thermo.service
I then ran the following commands:
sudo systemctl enable thermo
sudo systemctl start thermo
using ps aux | grep thermo I could see that the service was running:
pi@pi1:~ $ ps aux | grep thermo
root 901 2.4 4.4 88804 42096 ? Ssl 20:24 0:11 /usr/bin/python /home/pi/thermoPiDev/thermo/control/master.py
pi 1012 0.0 0.2 4276 2008 pts/0 S+ 20:32 0:00 grep --color=auto thermo
Finally, to test that the automatic restart worked, I used sudo pkill -g 901 -e to kill the process, and then ps aux | grep thermo to see if it
was restarted:
pi@pi1:~ $ ps aux | grep thermo
root 1057 53.6 4.2 87628 40576 ? Rsl 20:35 0:02 /usr/bin/python /home/pi/thermoPiDev/thermo/control/master.py
pi 1069 0.0 0.2 4276 1908 pts/0 S+ 20:35 0:00 grep --color=auto thermo
Note that the thermo process has a new PID, which I think means it was killed and successfully restarted.
Hopefully this will keep my pipes from freezing until I can actually fix whatever is causing the crash!