OTBR-Agent crashes with exit code 1, no other error

[aydinseven7]

I have an ESP32-H2 flashed with both esp-idf v6.0.1, v6.1.0 and v5.5.2 using the ot_rcp example (tried with all separately) + the openthread/border-router. Whenever I start the container, it crashes with exit code 1, but I dont see any other logs.

services:
  otbr:
    image: openthread/border-router:sha-85490d0
    network_mode: host
    privileged: true
    devices:
      - /dev/ttyACM0
      - /dev/net/tun
    volumes:
      - data:/data
    environment:
      OT_RCP_DEVICE: "spinel+hdlc+uart:///dev/ttyACM0?uart-baudrate=115200"
      OT_INFRA_IF: "eno1"
      OT_THREAD_IF: "wpan0"
      OT_LOG_LEVEL: "7"
      OT_REST_LISTEN_ADDR: "0.0.0.0"
      OT_VENDOR_NAME: "HomeAssistant"
      OT_VENDOR_MODEL: "OTBR"

volumes:
  data:

To Reproduce

1. Git commit id: I tried 85490d0, 78d9c28, c59a787
2. IEEE 802.15.4 hardware platform: ESP32-H2 with esp-idf 5.5-6.1
3. Network topology: OTBR container on server, ESP plugged into that server

Expected behavior
Container continuing to start like in the example. The tag used there is not available anymore though.

Console/log output

otbr-1 | s6-rc: info: service s6rc-oneshot-runner: starting
otbr-1 | s6-rc: info: service s6rc-oneshot-runner successfully started
otbr-1 | s6-rc: info: service fix-attrs: starting
otbr-1 | s6-rc: info: service fix-attrs successfully started
otbr-1 | s6-rc: info: service legacy-cont-init: starting
otbr-1 | s6-rc: info: service legacy-cont-init successfully started
otbr-1 | s6-rc: info: service otbr-agent: starting
otbr-1 | Configuring OpenThread firewall...
otbr-1 | Configuring OpenThread NAT64...
otbr-1 | Starting otbr-agent...
otbr-1 | [NOTE]-AGENT---: Running 0.3.0-85490d0
otbr-1 | [NOTE]-AGENT---: Thread version: 1.4.0
otbr-1 | [NOTE]-AGENT---: Thread interface: wpan0
otbr-1 | [NOTE]-AGENT---: Radio URL: spinel+hdlc+uart:///dev/ttyACM0?uart-baudrate=115200
otbr-1 | [NOTE]-AGENT---: Radio URL: trel://eno1
otbr-1 | [NOTE]-ILS-----: Infra link selected: eno1
otbr-1 | otbr-agent exited with code 1 (by signal 0).
otbr-1 | Chain OT_FORWARD_INGRESS (0 references)
otbr-1 | target prot opt source destination
otbr-1 | DROP all -- anywhere anywhere PKTTYPE = unicast
otbr-1 | DROP all -- anywhere anywhere match-set otbr-ingress-deny-src src
otbr-1 | ACCEPT all -- anywhere anywhere match-set otbr-ingress-allow-dst dst
otbr-1 | DROP all -- anywhere anywhere PKTTYPE = unicast
otbr-1 | ACCEPT all -- anywhere anywhere
otbr-1 | otbr-ingress-deny-src
otbr-1 | otbr-ingress-deny-src-swap
otbr-1 | otbr-ingress-allow-dst
otbr-1 | otbr-ingress-allow-dst-swap
otbr-1 | OpenThread firewall rules removed.
otbr-1 | s6-svlisten1: fatal: /run/s6-rc/servicedirs/otbr-agent failed permanently or its supervisor died
otbr-1 | s6-rc: warning: unable to start service otbr-agent: command exited 1
otbr-1 | s6-rc: info: service legacy-cont-init: stopping
otbr-1 | s6-rc: info: service legacy-cont-init successfully stopped
otbr-1 | s6-rc: info: service fix-attrs: stopping
otbr-1 | s6-rc: info: service fix-attrs successfully stopped
otbr-1 | s6-rc: info: service s6rc-oneshot-runner: stopping
otbr-1 | s6-rc: info: service s6rc-oneshot-runner successfully stopped

Additional context
Followed the guide at https://openthread.io/guides/border-router/build-docker

I have also tried adding different flags to the Radio URL:

• skip-rcp-compatibility-check
• different baudrate (H2 has 115200)
• vendor name is there because I got the error "Vendor not set" when using otbr-agent directly as entrypoint

[jwhui]

The error you are seeing typically indicates that otbr-agent is not able to communicate with the RCP.

Are you sure that the baud rate is correct? The Espressif docs indicate 460800.

https://docs.espressif.com/projects/esp-faq/en/latest/software-framework/thread.html

[chshu]

@aydinseven7 There are normally two USB ports on ESP32-H2 devkits, please check whether the ESP32-H2 is recognized as /dev/ttyACM0 or /dev/ttyUSBx?

For your reference: https://docs.espressif.com/projects/esp-thread-br/en/latest/qa.html#using-ot-rcp-with-otbr-ot-br-posix

[aydinseven7]

@chshu Both times it appears on /dev/ttyACM0, I have been trying with the USB port labeled "USB" instead of "UART". Flashing was done via JTAG, I guess that shouldnt have any impact though.
If the device is not mounted correctly, I get exit code 5 instead of 1

[chshu]

@aydinseven7 Which H2 hardware are you using, is it the official ESP32-H2 devkit?

If you are using the IDF ot_rcp example without any configuration changes, please use the UART port on the H2 devkit, which should be recognized as /dev/ttyUSBx.

The correct configuration should be:

OTBR_AGENT_OPTS="-I wpan0 -B wlan0 spinel+hdlc+uart:///dev/ttyUSB0?uart-baudrate=460800"

[aydinseven7]

@chshu Even when connected to the UART port, it shows up in ACM0. I was able to connect a screen session (got a blank display), but starting the border-router container resulted in this "output":

~��
~

So the connection is somehow used and screen does capture something, but I still get the same error message as before inside the container

[chshu]

@aydinseven7 Can you try enabling the OPENTHREAD_RCP_USB_SERIAL_JTAG configuration in ot_rcp, then connect through the USB port and try again?

If it still does not work, please build and flash the ot_cli example: ot_cli example. Then check whether the CLI commands work properly on your host. Let’s confirm whether the hardware connection path is working properly.