Postby NDJ_320 » Wed Feb 11, 2026 5:24 am
I am also trying to implement DFU in ESP BLE MESH, its successful(kind of) but there are some issues:
1. mesh_adv_task get overflowed when sending OTA binary to the unicast address after sometime of sending block
2. When group address is used, the device broadcasts to the group about the firmware metadata and cause stack-overflow on mesh_adv_task immediately after that
These are the major issues that are seen on the DFU, i am not sure why and also i am not able to debug the issue; nonetheless, the binary is being transmitted.
I have only tested on 2 servers, so i am not aware how many device the DFU can handle safely.
Also, why is it causing the stack overflow, i am not sure,. since all the data provided are persistent and already allocated before starting the mesh.
Sometime the sequence storage, etc.(operations that are performed when persistent connection is enable) also cause deep function calls that might cause stack-overflow. Likewise, disabling logs also helps to preserve stack, however the mesh_adv_task's watermark reduces again and again.