ltss broken since home assistant update to 2023.3.0 version about ltss HOT 13 CLOSED

v2.0.1 (proper release) is now available (https://github.com/freol35241/ltss/releases/tag/v2.0.1)

Thanks all for your help in finding this bug!

from ltss.

Do you have solved issue by restarting HA Core container ?

from ltss.

Hmm, I have the addon installed manually (not through HACS), I've removed the entire config/custom_components/ltss folder, downloaded the latest v2.0.1 release, and tried every method of restarting HASS (both with and without LTSS enabled in the config file), but any time I enable LTSS I still get the same issue:

Cannot quick reload all YAML configurations because the configuration is not valid: Component error: ltss - cannot import name 'with_metaclass' from 'sqlalchemy.util' (/usr/local/lib/python3.10/site-packages/sqlalchemy/util/__init__.py)

Do I need to manually forcibly refresh the dependencies somehow or something when not using HACS? If so, how?

Update: Figured it out. Renaming (or removing, if you're feeling brave) the config/deps folder and restarting will let Hass rebuild the dependencies, at least on hass.io/hassos.

from ltss.

Me too, but seems to work after restarting docker container manually.

from ltss.

Same error message, wount restart as long as i have ltss in configuration.yaml

from ltss.

That is interesting... I have now tried to reproduce this error in a clean environment (homeassistant==2023.3.0 and ltss==2.0.0) without success. It works flawlessly.

Can you all please provide some details about your setups? (Type of home assistant installation, preferably a list of the installed python packages).

from ltss.

I have the same issue and try to find the cause. It happens with the original and my custom / forked branch. First thought was my changes but it happens with the original too.

Unfortunately there isn't a lot more to go by. The error message is literally it. My current guess is the python compile cache but I'm neither sure if HASS has it enabled, nor where to find it.

I'm thinking about a caching issue because it tries to use Python 3.10 which now should be 3.11.

from ltss.

@Mornclaur you can force it to reboot now (depending on the device / installation) under the advanced options in the new dialog. It comes back up gracefully but fails to load the integration.

from ltss.

I am running HassOS 9.5 generic x86_64 on an intel NUC. As specified from HACS notification, I waited to get on last version of HA for making ltss component update to 2.0.0 version. It's after HA reboot, after update of ltss through HACS that I get the specified error message.

Regarding python version running inside HA Core container it's still python 3.10 from my side:

Use `ha` to access the Home Assistant CLI.
# docker exec -t -i homeassistant python -V
Python 3.10.10
#

See joined list of installed python packages inside my HA Core container.
python_packages.txt

from ltss.

Thank you @FredericGuardia ! That got me on the right track.

Please try the new pre-release v2.0.1-alpha (https://github.com/freol35241/ltss/releases/tag/v2.0.1-alpha) and report back.

from ltss.

Looks good so far 👍

from ltss.

I had the same error message. It's fixed in 2.0.1-alpha. Updating the requirement for GeoAlchemy2 did the trick. Thank you!

from ltss.

I confirm too. with this new release (2.0.1), I don't get anymre this error message. Thanks.

from ltss.