Hello,
I'm experiencing a disconnection problem on several BLE Mesh networks. My topology is 1 client with an average of 60 servers, for all networks. Client is connected to a host via serial interface model and it's also the network provisioner. Devices are based on nRF52832, clients have SDK Mesh 3.0.0 and SDK 15.3.0, servers have SDK Mesh 2.2.0 and SDK 15.0.0. Firmware is a kind of customization of genericOnOffSwitch, just to figure out the working principle of the system.
After few weeks of operations, I have 3 installations where all servers are unreachable. It seems that something goes wrong with the communications between client and servers. Host can communicate with the client ble chip, but messages sent to the server devices are not answered. Servers node are powered on and they are doing their own routines, so I guess they are not broken or in a trap (at least for the node working functions, I don't know if BLE mesh stack is working or stuck somewhere).
When the problem is present in a mesh, if I send to the client the command Remote Scan (using Model Specific Command), the client answer with a reboot message (0x81), I think it's rebooting the BLE mesh client. If I send any configuration message for example (I tried the Composition Data Get message using the Packet Send (0xab) message) the client answer a 0x8e error.
The only way I found to solve the issue, is to send the State Clear (0xac) command, to clean the client flash and after copy back the provisioning info and the address of the other nodes. After that commands sent to server to set or get something are working and also other commands are working.
I supposed the issue can involve replay cache index misalignment, I don't know if it could be the root issue. Any idea?
Thanks
