RFM69 id and group resets to i0 and g210

[rob570]

Hi,
I've a custom id and group for the RFM69 tx/rx. Don't understand why, at every boot I've to manually set the id and group to my configuration (id:15 and group:200) using ssh debugging and minicom.
Could someone help me to solve?
Many thanks!

[alexandrecuer]

Hello
What do you use, the receiver attached to gpio or an usb key ? You use the sniffer as an home assistant addon or with plain docker or an orchestration tool like portainer ?

[alexandrecuer]

This tool is very basic. I'm not that familiar with the jeelib format and sure it should be improved. Anyway it just reads the serial port and if "OK" is found, it tries to organize a MQTT payload :

emontx_sniffer/jeelib_sniffer.py

Line 110 in 7fe0dd0

if datas[0] == "OK" and len(datas)>=2:

[rob570]

Hi Alex,
I have an old RFM12pi (the one with ATMega, like in your documentation) connected to GPIO. I use the sniffer as an HASS addon, so to modify the config I had to enable ssh debugging and run your docker image.
I've solved adding a conf parameter and send it after the serial init.
It's rude but now it works!
Thank you for your code!

`RFM69_CONF = "15i 200g\n"
if VERBOSITY:
log.setLevel("DEBUG")
else:
log.setLevel("INFO")
info_message = f'MQTT_HOST : {MQTT_HOST} - MQTT_PORT : {MQTT_PORT}'
log.info(info_message)

def connect_to_serial(port, baudrate):
"""open a serial socket"""
try:
socket = serial.Serial(port, baudrate, timeout=0)
except (ValueError, serial.SerialException) as err:
error_message = f'error {err}'
log.error(error_message)
return None
socket.write(RFM69_CONF.encode('utf-8'))
success_message = f'connected to {port}@{baudrate}'
conf_message=f'Sent configuration:{RFM69_CONF}'
log.debug(conf_message)
log.debug(success_message)
return socket`

[alexandrecuer]

Nice job, I think i made tests with the RFM69 only, not the RFM12

I've seen your mods on your fork. Maybe we should consider integrate them in the repo, what do you think ?

The thing should be to add in the config.yaml some options like a SEND_CONFIG boolean set by default to false and a CONFIG_STR string set by default to your conf, so one can do or not all that job from the UI...

main...rob570:emontx_sniffer:main

just a question, why do you use your custom id and group ? security ?

[alexandrecuer]

Anyway, i thought that once you write to the rfm and save, you dont have you do it again...but maybe it is not the case with rfm12?

[rob570]

Hi Alex,
the fact my RFM12pi doesn't save the configuration sounds bad to mee too. Don't know if the eeprom is damaged.
I've done other changes, you can check my git:
added RFM12pi initialization string
added topic customization (I would like to push data on HASS, avoiding emoncms addon)
added ttyS0 port
Solved the deprecation warning con mqtt_client CallbackAPIVersion (upgraded to version 2)
if you believe they could be useful, please feel free to merge in your repo. I've done this only to learn git and docker starting from your code.
Thank you!
P.S. I've a warning on the RUN dockerfile command, after upgrading alpine I had to manage the python venv. My solution is not clean, but it works
P.P.S. I changed the id and the group trying to avoid noise caused by another transmitter in the neighbourhood

[alexandrecuer]

@rob570 : for the venv, i think you dont have to activate at startup, cf the CMD part of the dockerfile... You can just add this :

#!/ve/bin/python3

in the sniffer.py file

[alexandrecuer]

added topic customization (I would like to push data on HASS, avoiding emoncms addon)

Thats makes sense. Topic customization is a good addition. You want to use the mqtt broker addon and not the emoncms addon, isn'it ? You plan to use grafana or is hass storage enough for you ?

By the way, do you know there is an emonhub addon ?
https://hub.docker.com/r/alexjunk/emonhub

[rob570]

...

By the way, do you know there is an emonhub addon ? https://hub.docker.com/r/alexjunk/emonhub

Uh Doh! I didn't!
Anyway my Pi i s overloaded (I should buy the coral TPU for frigate) and another container could take it down.
To use you emonhub I should only create the config yaml isn't it? I'll try.
About the topic customization you're right, I would like to use the mqtt broker addon.
The warning raised after the build of my docker image is about the format on the CMD command, it is not a json array. Don't know why it works without enabling the venv, but it does.

[alexandrecuer]

To use you emonhub I should only create the config yaml isn't it? I'll try.

Yes. At least. A .devcontainer folder should help and will save your time so you can test everything with vscode...cf https://github.com/Open-Building-Management/emoncms

Dockerfile is here :

https://github.com/Open-Building-Management/containers/tree/main/emonhub

There is no CMD command (it was only intended to be run through docker compose or portainer. So it must be modified

The warning raised after the build of my docker image is about the format on the CMD command

Do you build locally ? Or use github actions as a pipeline ? With a recent version of docker, you should have no warning...

[rob570]

...

Do you build locally ? Or use github actions as a pipeline ? With a recent version of docker, you should have no warning...

I've built with:
docker buildx build --platform linux/arm/v7 -t rob570/emontx_sniffer:alpine3.21.2 https://github.com/rob570/emontx_sniffer.git

and the output was:
`1 warning found (use docker --debug to expand):

• JSONArgsRecommended: JSON arguments recommended for CMD to prevent unintended behavior related to OS signals (line 14)`

[alexandrecuer]

Did you test to build and run with the previous CMD ["python3", "jeelib_sniffer.py"]

with the #!/ve/bin/python3
at the top of the jeelib_sniffer.py file

[rob570]

Sorry I apologize for what I've written in a previous post. I had to activate the venv, that's why I've merged two command in the CMD line your syntax didn't work with my venv.

…

Il 16/01/2025 19:29 CET Alexandre CUER ***@***.***> ha scritto: Did you test to build and run with the previous CMD ["python3", "jeelib_sniffer.py"] with the #!/ve/bin/python3 at the top of the jeelib_sniffer.py file — Reply to this email directly, view it on GitHub #1 (comment), or unsubscribe https://github.com/notifications/unsubscribe-auth/APSNNL6KGDAWOCX2EKOXKLD2K7277AVCNFSM6AAAAABU6ZXYXKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKOJWGQZDAOBYHA. You are receiving this because you were mentioned.Message ID: ***@***.***>

[alexandrecuer]

You dont have to activate the venv.

Forget what i've said with the # shebang and check this :

https://github.com/Open-Building-Management/containers/blob/7d2c71d99d5ae3f0f2b7cc9d3c83c743427ca0c7/emonhub/Dockerfile#L59

If you do this, the path will know that you want to run the python version from the venv

You just have to add your CMD command after this ENV command

[rob570]

You just have to add your CMD command after this ENV command

Yes, it works! Thank you!