What are the customizations that should be done to Rasbian-lite for running home assistant to make it a bit more like Hassbian?
Now that hassbian is dead, I’d like to build a hassbian like install, but I’d prefer not to rediscover all the hassbian customizations one-by-one such as
The old Hassbian install doc mentioned a few things like
GPIO pins are ready to use.
Bluetooth is ready to use (supported models only, no Bluetooth LE).
In case anyone asks why I don’t want to use Hass.io (now called just Home Assistant): I’m building a new RPI 4, I don’t want to use Hass.io because I want regular, non-container host OS for flexibility. I’d like to be able to install packages with diagnostic and development tools, and things that access the hardware when I need them without having to build them in a container.
I’m going to try to use the “supervised” home assistant install – where the HA supervisor runs in a docker container and runs HA and add-ons in a container.
However, I’d like to be able to fall back to just a Python Virtual Environment.
Why not just start from the Hassbian image? I think the last version of the Hassbian image that got released was July 2019. The “RIP Hassbian” blog post said there would be a late Q4 2019 final release but I don’t think that happened.
Yes, if I were to only use hass.io in docker containers I don’t have worry about the rest of the host configuration.
On the other hand if I want to run the virtual environment, I do need to worry about the rest of the host configuration.
I think there are a bunch of config details I’m going to trip over. Raspbian lite seems to have just under 500 packages installed. My hassbian install has over 920.
One of the first things I found when I looked, hassbian’s /boot/cmdline.txt, adds dwc_otg.lpm_enable=0 to disable USB power management. Raspbian lite doesn’t do that.
I don’t know if that’s still required though seems like a good idea.
A number of reasons, I do a lot of work in Python. I’m comfortable with venvs. I’ve got some devices that have unsupported integrations. I’ve had to do some debugging and editing the code for those. Some of them depend on pypi modules that have bugs. etc.
I could probably exec into the container and do what I need to do there. I haven’t figured out how I’d cleanly pull in custom code and make sure it overrides modules that the system pulls in. I haven’t had the time to figure out how to setup and use a dev. env. for hass.io in containers.
Regardless, I think there are more customizations that hassbian used to do to the OS based on the experience that was gained from the years it existed. I’d rather not keep stubbing my toe rediscovering them.
FWIW, I’ve tried the “supervised install” method and I’m about to give up on it. I’m looking for some basic, though minimal information on managing the supervised installation: starting, stopping, restarting from the host system without having to reboot.
I can’t tell how to access the ha CLI commands, or whether those aren’t supported in the supervised install mode.
There must be some guidance someplace but I’ve yet to find it. I’ve started a new thread looking for docs. As best I can tell I’m getting answers that don’t apply to somewhat new “supervised install” method.
The easiest way is to install the Community version of the SSH & Web terminal add-on. Put it on a port other than 22 (I used the super-creative 23). Then ssh to that port and you will be “in” the supervised HA world. You will be able to start/stop/upgrade/etc. all from the command line.
the supervised install method is no where near new.
it’s just the name has changed recently. And you are likely experiencing the effects of that.
Hassio has been around for a couple of years.
it is now called “Home Assistant”
if you install it outside of a provided image (HassOS or some other complete OS image) and install it on a linux host on docker then it is now called “Home Assistant Supervised”.
If you install the non-hassio version of Home Assistant the you are running “Home Assistant Core”. And that can be installed in Docker or some other VM or directly in the OS via a virtual environment.