hass-gardena-smart-system icon indicating copy to clipboard operation
hass-gardena-smart-system copied to clipboard

Randomly stops updating?

Open antgustech opened this issue 1 year ago • 12 comments

Every now and then the device in the integration simply stops updating. Now for example no values have been updated for over 3 days.

Capture

Restarting the integration seems to resolve the issue but it comes back again.

antgustech avatar Jun 11 '24 17:06 antgustech

Same for me. I have Version 1.1.1-rc3

MarkusH1975 avatar Jun 25 '24 12:06 MarkusH1975

This depends on binary_sensor.gardena_smart_system_connection. This gets stuck each few days. Unfortunately the only workaround now is a restart. This can be automated.

Stoffl17 avatar Jun 26 '24 07:06 Stoffl17

What service are you using to restart the gardena integration? (or are you restarting the whole HA instance?)

tbnobody avatar Jul 02 '24 19:07 tbnobody

This branch seems to have a solution for the problem. Mine hasn't lost connection since I updated.

https://github.com/py-smart-gardena/hass-gardena-smart-system/pull/209

Fywqdk avatar Jul 08 '24 13:07 Fywqdk

This branch seems to have a solution for the problem. Mine hasn't lost connection since I updated.

#209

How do I change to this branch in HACS?

Bucky2k avatar Jul 09 '24 18:07 Bucky2k

Honestly I just copied the part of the code that differed from this branch into my integration manually. It is just a couple of lines of code

Fywqdk avatar Jul 09 '24 22:07 Fywqdk

Thanks, did the same via file editor. Curious if it will work

Bucky2k avatar Jul 10 '24 18:07 Bucky2k

Arh ok. I have it running in a docker container on my NAS so it's pretty easy for me to edit the files manually. That said; If you go to HACS -> Integrations -> Top right corner and click on the 3 dots -> User defined repositories (I think - it's translated to danish for me), then you can add other github repositories.  

Fywqdk avatar Jul 10 '24 21:07 Fywqdk

What service are you using to restart the gardena integration? (or are you restarting the whole HA instance?)

Yes, a full restart via homeassistant.restart

Last week, I also copied the few added code-lines from this repository here to my instance and situation seems to be improved.

Stoffl17 avatar Jul 12 '24 07:07 Stoffl17

I just released a new version 1.1.1-rc4 including all currently opened PR.

If this helps, please let me know.

grm avatar Sep 02 '24 20:09 grm

I just released a new version 1.1.1-rc4 including all currently opened PR.

If this helps, please let me know.

i just updated it, it took like 5 mins and it was down.. will try restart again to see if it keeps happening.. i think it was my network, so ignore above

basn avatar Sep 03 '24 05:09 basn

I just released a new version 1.1.1-rc4 including all currently opened PR.

If this helps, please let me know.

It is not so often is disconnecting right now, Trying establish connection and goes total disconnected maybe once in two weeks. Regarding network, it cannot be issue, is a rock solid. But another interesting thing is the pattern when is loosing connection. It loosing connection exact every 2 hours and restore after 10 seconds. It happening all the time, every 2 hours. Debug logs showing nothing interesting, Can it be related to this warning? WARNING (MainThread) [homeassistant.util.loop] Detected blocking call to load_verify_locations with args (<ssl.SSLContext object at 0x7f4ca119fd50>,) inside the event loop by custom integration 'gardena_smart_system' at custom_components/gardena_smart_system/__init__.py, line 86: await self.smart_system.authenticate() (offender: /usr/local/lib/python3.12/site-packages/httpx/_config.py, line 147: context.load_verify_locations(cafile=cafile)), please create a bug report at https://github.com/py-smart-gardena/hass-gardena-smart-system/issues

jires avatar Oct 06 '24 19:10 jires

@jires : I think the warnning refers to another issue, that should be fixed in the next release. I jsut updated the mainstream library. The connection is probably related to token expiration and the fact that we have to get another token. I'll close this issue for now, if the problem occurs again, feel free to create another one :)

grm avatar Jun 01 '25 14:06 grm