Gateway stops sending data for devices after running for a period of time. about iotedge-lorawan-starterkit

[ronniesa]

One trick that you can do is to change a dummy env variable on each module this will trigger a restart of every module. Maybe is worth to try just forcing the restart of the edgeHub module that could be the cause of not correctly reconnecting.

If you have a lot of network disconnects is also recommended to put a DNS settings in to edge: https://docs.microsoft.com/en-us/azure/iot-edge/troubleshoot#edge-agent-module-continually-reports-empty-config-file-and-no-modules-start-on-the-device

Yes, it would be nice to be able to force a reboot of edge from IoT Hub.
You could implement a module that runs in priviledged mode and allow you to call a directmethod to trigger the reboot. Not probably a best practice but it should work.

[gerfen]

@ronniesa Thanks for the suggestions. I think I will start with the DNS settings to see if that improves the situation.

[Mandur]

Hello @gerfen, I had luckily a similar requirement for our ci. Sadly my method don't work on x86, but I found a more aggressive one that can work with x86 devices : https://github.com/Mandur/IoTEdgeDeviceRebooter.

[ronniesa]

@gerfen did the DNS entry solve the issue?

[gerfen]

@ronniesa Unfortunately I did not have time to try to add the DNS entry while I was on site as I got side tracked with not be able to join my new devices to the gateway. I have plans to return in two weeks and will attempt to add the DNS entry then.

I can close this issue and re-open i if necessary.

[ronniesa]

@gerfen let's leave it open for now until you test it out so we can better track it

[AxelBertilsson]

@gerfen Any progress on the issue?

[gerfen]

Unfortunately no due to the current situation here in the States. I will hopefully be back on site in two weeks. We can close the issue and reopen it if appropriate.

[Mandur]

@gerfen Were you able to test the proposed solution?

[gerfen]

@Mandur No I haven't been able to test the solution. That said the system has stabilized. The real issue was related to the max number of devices getting reset to 100 when I did an upgrade. It took me ages to figure that out. I'll close this issue now. Sorry for the fire drill.