openthread cli does not auto start and cannot use commissionner

[wass]

I'm trying to create an openthread network extension with nrf52840 dongle. For that, it needs to be Full Thtead Device so that it's capable of becoming a router and should be able to join automatically so auto start thread and joined config. I based it on the cli example

github.com/.../cli

witht the following west.yml

https://github.com/HomeSmartMesh/sdk-hsm-thingy53/blob/main/west.yml

which has nrfconnect v2.3.0 and sdk zephyr v3.2.99-ncs2

with the following additional overlay

github.com/.../overlay-joiner-rounter.conf

here the content for conevenience 

```

CONFIG_OPENTHREAD_NORDIC_LIBRARY_FTD=y
CONFIG_OPENTHREAD_FTD=y
CONFIG_OPENTHREAD_MTD=n
CONFIG_OPENTHREAD_MANUAL_START=n

# OpenThread joiner
CONFIG_OPENTHREAD_JOINER=y
CONFIG_OPENTHREAD_JOINER_AUTOSTART=y
CONFIG_OPENTHREAD_JOINER_PSKD="ABCDE2"
CONFIG_OPENTHREAD_SLAAC=y

CONFIG_OPENTHREAD_COMMISSIONER=y

```

Yet, by default the state is disabled and is not autostarting despite `CONFIG_OPENTHREAD_MANUAL_START=n`

plus the commissionner command does not exist, I get an error when I try to use it on the shell, despite `CONFIG_OPENTHREAD_COMMISSIONER=y` 

I built it as follows and checked in `build/zephyr/.conf` that the config flags are taken correctly. 

`west build -p always -b nrf52840dongle_nrf52840 -- "-DOVERLAY_CONFIG=overlay-usb.conf;overlay-joiner-rounter.conf" -DDTC_OVERLAY_FILE="usb.overlay"`

I also perform an `ot factoryreset` for every test and have a sniffer to check packets. Right after a factory rest, and after the dongle restarts and the serial attaches again, the `ot state` is disabled and not detached or other, so I have to manually run`ot thread start` which is a manual start and not automatic.

My end goal is to make MTD devices be able to join through a cli dongle as router, but for that I assumed the cli has to be capable of commissioning first, and failed in testing a simple cli leader that can commission from its shell.

[wass]

I answer part of my own issue, as I clarified the "commissioning" part. It was not clear to me when a dataset is needed and when not.
when testing a new fresh factoryreset cli and creating a dataset as described here, it works

https://developer.nordicsemi.com/nRF_Connect_SDK/doc/latest/nrf/protocols/thread/overview/commissioning.html#setting-up-the-commissioner

I'll still investigate why it does not autostart, and when it comes to other issues (joining from intermediate router, I'll test more and might open an issue in a different ticket).

[AHaug]

Hi,

Glad to hear that you've solved (atleast partially?) the issue you were observing. Could you verify if you've solved it or not?

If not, I'll continue supporting you with solving the case! 

Kind regards,
Andreas

[wass]

Hi,
thank you for taking this up.

I'll try to simplify and reduce the scope of the question.

With the above joiner config applied on the cli sample, is the following scenario expected 

uart:~$ ot factoryreset
------reboot
uart:~$ ot state
disabled
Done

As you can see, the cli sample starts as disabled and does not attempt to join any network as per the expected `CONFIG_OPENTHREAD_JOINER_AUTOSTART=y`

If that is somehow expected, that would answer the question as it is different than normal samples, the cli is maybe overriding the default config ?

[AHaug]

Hi,

From what I know it should autojoin. Could you have a look at this git issue from zephyr and check the logs to see if the issue is similar as the one reported in the issue and try the suggested fix: github.com/.../30433

It seems that you have not erased the OT settings area on flash (west flash --erase). Looking at the logs, your device already had a dataset stored and it simply formed a network. The startup code won't bother with the joining procedure if it has a valid dataset stored in flash.

If it does not solve the issue, could you check and see if it is reproducable in a somewhat minimal sample such as https://github.com/hellesvik-nordic/samples_for_nrf_connect_sdk/tree/main/openthread_samples/autojoin? 

Kind regards,
Andreas

[wass]

Hi,

thank you for the suggestions. I checked the github issue link, that is something I was aware of and is quite common for newly flashed devices in order to be in a factory reset state right after flashing. Flashing actually does not delete the openthread settings region, that's why, what actually deletes the openthread settings region is calling "ot factoryreset" which I included in my test as shown above.

I think the question could be answered by investigating the code or having a clear overview on what the cli does, which is a bit challenging to me. If the code proves that the behavior is expected then reproducing it will not justify if it is a bug or not, maybe the cli example was just not intended to be used like that.

In order to make reproduction easier, as requested here I put together in one directory all what is needed, the readme also includes the needed command which must include the test overlay

https://github.com/HomeSmartMesh/sdk-hsm-thingy53/tree/main/samples/30_usb_dongle_router

here the commands from the readme

```shell

west build -p always -b nrf52840dongle_nrf52840 -- "-DOVERLAY_CONFIG=overlay-usb.conf;overlay-joiner-rounter.conf" -DDTC_OVERLAY_FILE="usb.overlay"

nrfutil pkg generate --hw-version 52 --sd-req=0x00 --application build/zephyr/zephyr.hex --application-version 1 build/zephyr/zephyr.zip

nrfutil dfu usb-serial -pkg build/zephyr/zephyr.zip -p COM9

```