Right - so I’ve been battling with this, on and off, for months - but it’s got a lot more frequent recently. I did think it might be aligned to the echo binding memory leak, but since that’s been resolved - I know it’s not that.
So - the problem - every so often - the RFXCOM binding just stops receiving/transmitting. No errors in logs, it just stops. Restarting the binding doesn’t fix it, but a full restart of Openhab will. Openhab otherwise continues on quite happily.
Running this on a RPI 3B with a RFX433E (about 2 years old) running the latest Pro1 firmware, running raspbian with openhab installed on top. Been using the same setup for about the same amount of time, but have upgraded underlying OS and Openhab consistently. Running 2.5.5 with the amazon echo binding up to the fix provided in the memory leak/connection problem thread. The 433e is monitoring a weather station, an energy monitor and is controlling 2 Somfy blinds which are only activated at sunrise and sunset.
There’s no consistency in timing, events or otherwise. I’m finally posting this with some logs on trace level to see if that provides anything.
So - here’s the situation with this afternoon:
14:58 - last message recieved by the RFXCOM binding
transceiver checks continue and report the device online.
Around 15:55 I notice the binding has stopped
15:58 - I attempt a restart of the binding - this disposes of the various items, but disposing of the handler takes too long (apparently). After that - the tranceiver is unitialised.
Around 2 minutes later - I attempt a USB hot port swap for the transceiver to see if this resolves the issue, but the transceiver doesn’t come back online in OpenHab. (note - I’ve done hot swaps when it IS working with no issue previously)
So I then restart OpenHab (note - I didn’t restart the PI) and it comes back absolutely fine.
Now previously when this has occurred, doing things like a USB port swap can seem to resolve the problem for a month or two, but the issue has always returned. It’s possible I’ve had multiple issues over time of course and swapping port resolved problem A - but not problem B.
Following here: RFXCOM Binding stops working after some time - I also made sure the 433e was on the latest firmware (it wasn’t two weeks ago but it is now) - but it’s still suffering.
I don’t quite believe it’s related to the device because it doesn’t require a full reboot to restore the transceiver to working - it just needs openhab to be restarted.
What’s notable about where it stops is that, in the terms of the logging, it’s made to to RFXComBridgeHandler:237, but never makes it to RFXComHandler:146 - the normal path you see from other events and I just can’t see how that can occur, but then I’m really not deep into the flow of how this binding all hangs together.
Anyway - logs attached - I’ve trimmed from before the event - but included enough to show the same rain device has reported before successfully.
Any thoughts on next steps on how to diagnose would be very welcome.
I’m about to add more blinds to the system so I’m very keen to get this resolved. Having the kitchen blinds not open/close automatically isn’t so much of a pain if I haven’t spotted the hang, if it’s going to affect more rooms, that’s a problem.