Running post_to_influxdb.py for a long period creates hundreds of zombie hcitool processes

[zbrozek]

It also stops actually capturing and uploading data once this starts happening.

root 28948 0.0 0.3 7220 3216 pts/1 S+ Apr19 0:00 sudo -n hcitool lescan --duplicates
root 28955 0.1 0.0 2476 632 pts/1 S+ Apr19 11:05 hcitool lescan --duplicates
root 29302 0.0 0.3 7220 3420 pts/1 S+ Apr19 0:00 sudo -n hcitool lescan --duplicates
root 29308 0.1 0.0 2476 624 pts/1 S+ Apr19 11:22 hcitool lescan --duplicates
root 29384 0.0 0.3 7220 3332 pts/1 S+ Apr20 0:00 sudo -n hcitool lescan --duplicates
root 29391 0.1 0.0 2476 648 pts/1 S+ Apr20 10:29 hcitool lescan --duplicates
root 29406 0.0 0.3 7220 3332 pts/1 S+ Apr20 0:00 sudo -n hcitool lescan --duplicates
root 29413 0.1 0.0 2476 616 pts/1 S+ Apr20 8:43 hcitool lescan --duplicates
root 29642 0.0 0.3 7220 3416 pts/1 S+ Apr19 0:00 sudo -n hcitool lescan --duplicates
root 29649 0.1 0.0 2476 624 pts/1 S+ Apr19 11:27 hcitool lescan --duplicates
root 30432 0.0 0.3 7220 3416 pts/1 S+ Apr20 0:00 sudo -n hcitool lescan --duplicates
root 30438 0.1 0.0 2476 668 pts/1 S+ Apr20 9:08 hcitool lescan --duplicates
root 30466 0.0 0.3 7220 3216 pts/1 S+ Apr20 0:00 sudo -n hcitool lescan --duplicates
root 30473 0.1 0.0 2476 656 pts/1 S+ Apr20 8:37 hcitool lescan --duplicates
root 30576 0.0 0.3 7220 3320 pts/1 S+ Apr20 0:00 sudo -n hcitool lescan --duplicates
root 30583 0.1 0.0 2476 624 pts/1 S+ Apr20 9:12 hcitool lescan --duplicates
root 30700 0.0 0.3 7220 3420 pts/1 S+ Apr20 0:00 sudo -n hcitool lescan --duplicates
root 30707 0.1 0.0 2476 636 pts/1 S+ Apr20 9:02 hcitool lescan --duplicates
root 31163 0.0 0.3 7220 3436 pts/1 S+ Apr19 0:00 sudo -n hcitool lescan --duplicates
root 31171 0.1 0.0 2476 656 pts/1 S+ Apr19 11:03 hcitool lescan --duplicates
root 31322 0.0 0.3 7220 3348 pts/1 S+ Apr20 0:00 sudo -n hcitool lescan --duplicates
root 31329 0.1 0.0 2476 616 pts/1 S+ Apr20 10:39 hcitool lescan --duplicates
root 31853 0.0 0.3 7220 3436 pts/1 S+ Apr23 0:00 sudo -n hcitool lescan --duplicates
root 31860 0.0 0.0 2476 624 pts/1 S+ Apr23 0:14 hcitool lescan --duplicates
root 31930 0.0 0.3 7220 3320 pts/1 S+ Apr21 0:00 sudo -n hcitool lescan --duplicates
root 31935 0.1 0.0 2476 620 pts/1 S+ Apr21 8:08 hcitool lescan --duplicates
root 31998 0.0 0.3 7220 3204 pts/1 S+ Apr23 0:00 sudo -n hcitool lescan --duplicates
root 32005 0.1 0.0 2476 668 pts/1 S+ Apr23 2:10 hcitool lescan --duplicates
root 32276 0.0 0.3 7220 3364 pts/1 S+ Apr21 0:00 sudo -n hcitool lescan --duplicates
root 32283 0.1 0.0 2476 620 pts/1 S+ Apr21 6:07 hcitool lescan --duplicates
root 32712 0.0 0.3 7220 3340 pts/1 S+ Apr21 0:00 sudo -n hcitool lescan --duplicates
root 32718 0.1 0.0 2476 660 pts/1 S+ Apr21 8:25 hcitool lescan --duplicates

[ttu]

There must be something wrong with killing hcitool process
https://github.com/ttu/ruuvitag-sensor/blob/master/ruuvitag_sensor/ble_communication.py#L90

Any ideas?

[zbrozek]

Possibly? But how would I find out? It doesn't look like errors are propagated anywhere. Maybe the stored PID is wrong?

[ttu]

Will make a fix for hcitool process killing

[ttu]

Fixed hcitool subprocess closing (5deed32).

Also updated the https://github.com/ttu/ruuvitag-sensor/blob/master/examples/post_to_influxdb.py code. For long running processes should use RuuvitagSensors.get_datas-method as it won't start a new BLE scanning subprocess with each method call.