Sure, I get it. But my homeassistant machine is also my HTPC. And I had issues with some packages available on Debian. That is my least effort way of dealing with it.
I’ve just reversed engineered supervised installer and converted to arch pkg. Also, any help to maintain is appreciated.
I already had frigate installed as docker container, and tuned /etc/docker/daemon.json for SHM upgrade. The AUR package complained the file is already present on the FS.
I backed it up away, installed homeassistant-supervised and merged my setting manually.
I suggest a more flexible way of installation of the file, but I am not competent on Arch (yet). On debian I know there are a bunch of package maintainer script to merge configuration files and not ship them “barely” as normal package files.
I think there is the same thing on Arch but don’t know how it is called nor how it works.
Just to let you know, as far as I configured it (and after having done all the apparmor stuff you wrote), your supervised version works like a charm.
I’m currently looking for a “clean” way to bind-mount read-only the frigate host storage path, just for HA to display the remaining storage in my dashboard via the sensor.systemmonitor. I already got /, I want /home/data.
Where is the “main” HA container configuration file ?
EDIT: I will start here : [Solved] Hassio and volume mount issues - #2 by febalci This doesn’t seem very clean because changes are not kept when container is updated, I have to redo the configuration. Is there any way to have a docker-compose.yml file somewhere ?
The files are located at: /usr/share/hassio/homeassistant/. Just add a link inside that path and it’ll do the job. Remember to add the right permission to that folder.
For anyone landing here and seeing their supervisor reported as “Unhealthy” even after following all the installation steps listed in the Arch Wiki page, try this:
Triple-check you have apparmor service running
Make sure you have added the kernel parameters (see previous comments)
Re-install the homeassistant-supervised AUR package
This made my supervisor “Healthy” for a brief moment. Then I checked the supervisor logs:
docker logs -f hassio_supervisor
I noticed it was being flagged as “Unhealthy” due to running the watchtower container on the same host. This seems to make the supervisor unhealthy. I assume an incorrectly configured watchtower container could mess up with the running Hassio containers so they just flag it.
After stopping and removing watchtower, it shows as “Healthy” and I can install add-ons as expected.
@adi-mistrzu, thanks for the suggestions. The first one was fixed, although both will work the same way. The second is incorect. MACHINE should be one of the following:
generic-x86-64
odroid-c2
odroid-n2
odroid-xu
qemuarm
qemuarm-64
qemux86
qemux86-64
raspberrypi
raspberrypi2
raspberrypi3
raspberrypi4
raspberrypi3-64
raspberrypi4-64
tinker
khadas-vim3
That’s why we have a check_machine function in the install script. If the CPU arch isn’t i386, i686 or x86_64, you need to set MACHINE env var.
Now, about your error, looks like a network configuration issue. Check if NetworkManager is correctly installed and configured.
Also, check System Health - Home Assistant
Should look something like this:
I managed to run supervisor once, I suppose there is a problem with IP addressing
Could you show me you ifconfig ? docker and hassio shoudn’t be in the same network?