OTBR lost some frames when sending data to multiple end devices at the same time

[Semi-Bin-Blain]

Hi there, I find that the OTBR lost some frames when it transmits the data from the Cloud Platform to the multiple end devices at the same time.

I have 5 End devices whose Thread Roles are all MED, these End Devices communicate with the Cloud Platform via COAP protocol. I can send commands from the Cloud Platform to the end devices; I sent the same command to the 5 devices at the same time, but not all the devices can receive the command every time, so I caught the network frames in the OTBR wpan0 interface and caught the Thread frames via the sniffer, the result shows that the 5 frames which include the command have arrived the OTBR wpan0 interface, but only 4 of them sent to the end devices by the OTBR via the OpenThread RCP. (like the picture shows)

wpan0 frames:

sniffer frames:

The information as follow:

1. Git commit id: f0bd2166467464c3fac23b28f7669179ee9ba332
2. IEEE 802.15.4 hardware platform: NRF52833DK
3. Thread Version 1.1
4. RCP software: ncs1.8.0/nrf/samples/openthread/coprocessor (ncs1.8.0)
5. OpwnWrt Version: stable release 21.02
6. the Thread network key(Master key): 93864f17ccd7f13f1834a738907994f6
7. the .pcapng files: pcapng_files.zip

I am not sure what is causing the problem, so I come for help.

Thanks in advance

[jwhui]

@Semi-Bin-Blain , one thing to check is whether OpenThread is dropping any messages.

We can use the counters CLI command to gain more insight. For example, can you provide the output of the following commands?

• counters mac
• counters mle
• counters ip

[Semi-Bin-Blain]

@jwhui Thanks for your reply, I've tried again, and the output is as follows file. I got the log only when the problem happened, and I noticed that the ”counters IP“ can directly show that the problem happened, the “Tx failed” of “counters IP” would increase by 1, sometimes, so did the "Rx failed" ethier.

OT_CLI_counters.log

[jwhui]

Closing stale issue.