It turned out to be too much hassle, so I gave up running hass core and moved to container based deployments. Most of the work was to give the container access to the SkyConnect stick and the fail2ban log. Hereâs my Docker compose file:
version: '3.7'
services:
home-assistant:
container_name: home-assistant
image: homeassistant/home-assistant:2024.12.3
restart: always
user: 1000:1000 # UID and GID of the 'pi' user on the Raspberry Pi host
group_add:
- 4 # 'adm' group for fail2ban log
- 20 # 'dialout' group for SkyConnect stick
environment:
- TZ=Europe/Amsterdam
devices:
- '/dev/serial/by-id/usb-Nabu_Casa_Home_Assistant_Connect_ZBT-1_fcbd3b7b7b31ef11985b51cfdfbc56eb-if00-port0:/dev/serial/by-id/usb-
Nabu_Casa_Home_Assistant_Connect_ZBT-1_fcbd3b7b7b31ef11985b51cfdfbc56eb-if00-port0'
# target should be something like /dev/skyconnect for new installs, but this is
# a migration so keep the old name
network_mode: host
volumes:
- type: bind
source: /home/pi/.homeassistant
target: /config
- type: bind
source: /var/log/fail2ban.log
target: /var/log/fail2ban.log
read_only: true
Right now I have also included the zwave-js-ui service in the same Docker compose file and made the home-assistant container depend on its health, but this dependency could also be made in the systemd unit files.
Shelly dimmer 2 modules turn on but one canât then turn them off afterwards. Same thing with them being on initially, in that one then canât turn them off.
Iâd skipped the first release of 2024.12, where 2024.12.2 and 2024.12.3 worked perfectly but 2024.12.4 exhibits this problem.
checking the display again, I now noticed a flashing wifi icon.
which is somewhat puzzling as the router shows the device being online just fine (and its not the connection to the outside station, because that has its own connection icon)
every data point was fine, so all I could think if was to unplug the power cord (it has batteries also âŚ) wait 10 sec and reconnect.
lo and behold the entities popped back to life.
the mysteries of Home automation⌠was it all a coincidence? guess I will never know
I upgraded to version 2024.12.4 and am having issues with the Roborock integration. I can log in, receive a 6-digit code from Roborock, enter the code into the integration and it appears everything is good except the integration shows âFailed to Setupâ.
Here is a snippet from the logs:
Logger: homeassistant.config_entries
Source: config_entries.py:640
First occurred: 1:10:01 PM (9 occurrences)
Last logged: 1:37:58 PM
Error setting up entry [email protected] for roborock
Traceback (most recent call last):
File â/usr/src/homeassistant/homeassistant/config_entries.pyâ, line 640, in __async_setup_with_context
result = await component.async_setup_entry(hass, self)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File â/usr/src/homeassistant/homeassistant/components/roborock/init.pyâ, line 72, in async_setup_entry
device.duid: device for device in all_devices
^^^^^^^^^^^
AttributeError: âdictâ object has no attribute âduidâ
Unfortunately, nothing is appearing in my logs in regard to this issue. What I have discovered is that after the update to 2024.12, Perform action âSwitch: Turn onâ on Light Group (LR) - Switches doesnât work anymore, but Switch âTurn onâ on Light Group (LR) - Switches does work in my automations
Didnât think about posting them. But I did turn on debug logging so Iâll be sure to post soon. Thanks for the reminder.
Could have been kinder, but a good point nonetheless.
@jerryinc@EricT Your issue was due to an api change and has been resolved with the next update
hereâs extra context from another thread:
Hi everyone! Codeowner for the Roborock integration here. I have a fix for this already included in 2024.12.5 as of this morning. Once 2024.12.5 gets pushed out, update to that.
Basically, Roborock changed their api yesterday a tiny bit and that is causing the issues. I added a solution that will help prevent a similar error from happening again, but api changes of any form can cause major problems.
MANY previous versions of the roborock integration will be permanently broken! So Iâm sorry to force anyone to 2024.12.5 that wasnât planning on it, but you will need to. I encourage everyone to go here: #133533 where I will leave the issue open for any further questions. I do not frequently check the forums.
Sorry for the inconvenience everyone! I fixed it as soon as I was aware of it, but of course I have no real affiliation with Roborock, so I did not know beforehand. Hopefully 2024.12.5 will be released shortly.
Is there something specific I have to do to enable the fall back option for assist? I donât see the option to enable âprefer handling commands locallyâ in my 2024.12.4 containerized version.
Just a callout for everybody using HA Core directly (yes, I know weâre a minority and responsible for dependencies ): I noticed an error I didnât see until I upgraded from 2024.11.3 to 2024.12.4.
[aiohttp_fast_zlib] zlib_ng and isal are not available, falling back to zlib, performance will be degraded.
You need to enable the isal integration (assuming you have the required dependencies installed). This is new since 2024.6 and isnât included in the default config.
Configuration of this integration only applies to Home Assistant Core installations types. Home Assistant Container, Home Assistant Supervisor, and Home Assistant Operation System installs already have isal pre-installed, and no action is required.