Hello everyone,
I'm encountering a persistent issue with several NKE Modbus Class C sensors that abruptly disconnect from my ChirpStack environment and subsequently fail to re-establish a connection successfully. I'm hoping someone in the community has experienced and ideally resolved a similar scenario.
My setup is as follows:
- ChirpStack Version: 4.14.1 (Ubuntu Linux)
- Gateways: 2 x Dragino LPSN8 (lgw-5.4.1753346938)
- Device Type: NKE Modbus Sensors Class C
- Activation: OTAA
- LoRaWAN Specification: v1.0.2 (NKE recommendations)
- Region: Region B
- Class C
- Confirmed downlink timeout set to 5 seconds
Basically, I have some Modbus Sensors that suddenly disconnect not necessarily at once and sometimes one of them reconnects on its own ... But Following the disconnection , the devices attempt rejoin, sending join request and I see sometimes join accept but no uplinks just another join. If i see a DevNonce error I try to flush the device nonces.
Sometimes if I restart chirpstack, redis-server and the two gateways the sensor reconnects magically at the next join but it doesn't work all the time and its not a reliable fix.
I tried to check the logs but there are no explicit error messages that could lead me the cause. I noticed lots of mac commands going so I decided to change the ADR algorithm for something more stable but I doubt this would work.
Has anyone else experienced NKE Modbus Class C sensors, or any Class C devices, showing this pattern of sudden disconnection followed by repeated failed join attempts (many join but no uplinks, or DevNonce Error)?
Any suggestions on where to focus my debugging efforts, gateway configuration details, or known interactions between NKE devices and ChirpStack v4.x would be greatly appreciated !
Thank you in advance for your insights !
Here are some logs I retrieved in case it rings a bell :
https://pastebin.com/wvy0mzSi