lm-sensors is daemonized and run as a service. When Supervisor starts lm-sensors, it thinks it exits immediately, so it retries a few times before giving up. This is harmless, as the service starts the first time.

However, there are erroneous log messages in syslog as follows:

INFO spawned: 'lm-sensors' with pid 20
INFO exited: lm-sensors (exit status 0; not expected)
INFO spawned: 'lm-sensors' with pid 29
INFO exited: lm-sensors (exit status 0; not expected)
INFO spawned: 'lm-sensors' with pid 44
INFO exited: lm-sensors (exit status 0; not expected)
INFO spawned: 'lm-sensors' with pid 58
INFO exited: lm-sensors (exit status 0; not expected)
INFO gave up: lm-sensors entered FATAL state, too many start retries too quickly

Also. supervisor cannot watch it properly, and thus cannot react if lm-sensors ever terminates.

Need to find a way to run lm-sensors in the foreground.