Nova timeout mystery...



My setup is a headless GPS tracker, based on Pi 3 with a Nova LTE. Reading Ben Strahan somewhere he prefers systemd for start-on-boot, I have a systemd script that does the following (on boot):

  1. connect the Nova
  2. check for updated scripts (on G Drive, with rclone) and update if necessary
  3. start a TCP server (to take commands like START, STOP, REBOOT, etc. from Hologram web-based TCP client)
  4. autostart a first session (a “session” is the script that handles the GPS receiver data and publishes info to Thingsboard).

So, when powering up the Pi, after about a minute, I can see it start publishing it’s location, with my defined intervals, to Thingsboard. Everything smooth so far.

The weird thing is that it stops publishing after 1 minute. Looks like the Nova connection dropped. Onboard logging continues so the GPS module is working fine.

Now also, it can NOT be the Nova, because the Nova still takes commands on the TCP socket! So the Nova is alive! Unless it gets woken up by the web-based TCP client…

I thought it might be a Thingsboard-related session timeout, but this is not the case, because when I connect the Pi to internet over my mobile phone hotspot (via wifi), publishing continues for as long as I want.

Hence my conclusion that there’s a Nova-related 60 second timeout somewhere, but I wouldn’t know where…

EDIT: when I manually STOP the session and START a new one, again it publishes for 60 seconds only…


Interesting…I am seeing similar behavior although I haven’t keyed in on a 60 second interval. I’ve had it stop quicker than that as well as persist longer than that.