God yet another thread in the same sphere.
That’s making sense.
(I’m slowly getting this guys, thanks).
So I’m back to re-reading the installation (https://www.home-assistant.io/getting-started/) page with a newfound knowledge that half the videos on the internet about Home Assistant are talking about Home Assistant Core because back then, that’s what it was called, and not the Home Assistant of today, which is…
The penny just dropped.
-
Home Assistant (of today) IS (just) an image, designed for specific hardware (Pi, Tinkerboard, NUC etc) or a VM and is “complete”. This is most akin to what was once called hass,io. (hence hass,io >> HA)
-
Home Assistant (that is in most of the YT videos) is what is now called Home Assistant - Core, and
-
a) Home Assistant - Core (as a python package on “whatever”) can still have the components of Supervisor and Supervisor-managed “add-on’s” OR *
-
b) Home Assistant - Core (as a python package) can include bare-metal self managed components
* is the only one I’m not sure about now.
Yes you seem to be understanding now!
So we say the supervisor is the docker manager, than the addon are the docker containers specifically made to be run by the supervisor – and HassOS is the underlying operating system of it all. When the container is an “addon”, it is managed by the supervisor, which allows it to be configured from the HA UI.
If you don’t have the supervisor, then you can’t use add-ons, instead you would use a plain docker container – For example many people run Plex media server, which is also available as an addon – but what if you didn’t use HA – How would you install Plex then – point is, anything that is an “add-on” can still be installed and configured separate but still work with HA.
Well, if the front page of the Home Assistant website had its own cheat sheet explaining the evolution of the product, then people wouldn’t be so confused. By the sound of your post, it’s a regular occurrence and (as stated before) given the fact that information on the web dates very quickly, someone’s explanation of it last year is not necessarily valid or accurate today.
Hence the need for another one of these threads.
It’s hard to get… I had to read numerous explanation, all saying the same thing, said by different people before I finally “got it” – And it’s easier for some of us that have been around awhile, only having to learn one new option at a time as opposed to jumping in now when there’s multiple ways to run HA
I am not criticising you, it just gets boring that there is no moderation to speak of on this forum. If there was, you’d have been referred to the latest word and then this thread locked, or merged to the blog thread.
I think this seemed ( mostly ) more about trying to figure out what’s what. Not necessarily trying to reopen that previous debate
It explains what is what.
Well, we also rely on people flagging things if they think mod attention is required Seriously… people have to flag stuff, or it relies on mods spotting a thread and happening to think the same way.
At which point the pitchfork brigade comes out again, complaining about censorship
Yeah ok, it’s a fair post, but I have got more from this thread than that. I haven’t read the user comments yet, I’m not at home and I’m on mobile but wanted to quickly reply.
From what is in that blog post I would have been a little unclear of the subtly between what’s described as “Home Assistant” vs “Home Assistant Supervised”. Notably:
This method was previously known as “Hass,io”, and includes our Operating System (HassOS), the Supervisor, and add-ons. It can be run on various single-board computers or in a virtual machine.
As Supervised is described the same way, both previously known as “Hass,io” and both running on Linux computers. The difference being that the first includes HassOS.
I get it now, especially after the help from here. So I thank people like you for allowing threads like this to stay open and allow noobs like us to get a chance to understand it. I’m glad That people took the time and it’s greatly appreciated.
Thanks
But can’t everybody just post something on that GitHub discussion?
Also I think the problem is partly our own fault (yes, re-using HA wasn’t clever either) because a lot of us still use hass.io instead of HA.
Just adding my 2 cents here (or 1) - to me, the name hass.io has been pretty strange, but easy to use. However, my German girlfriend has a bitter laugh, since ‘Hass’ means hate in German and is a word associated mostly with right-wing things. I’m not a political guy, could care less, but I think it should still be avoided.
By extension, ‘hassbian’ sounds like a hate for *bians, whatever the star signifies.
This post is a useless contribution and mostly a curiosity.
It’s not useless - it’s exactly why Hass.io was dropped, though Hass is still in use for things like HassOS, and is an easy shorthand for English speakers
I really do think I get most of it, it’s just the “docker” part of supervisor that I want more clarification on before I decide to (most likely) go down that path.
Does the supervisor manage the updates of the components that make up the add-on’s AS WELL AS the docker container system itself? Or is docker maintenance up to the user?
The reason I ask this is because a) I am comfortable with a complete OS, I’m happy with VMs, in fact I much prefer them. But docker containers and getting them to play nicely with each other has always been beyond me, and although I like the complete “supervised” side of it, I’m hoping that the supervisor takes care of “everything docker” (which I’m assuming, it must).
@Tinkerer and @frits1980 I’ve read through half of the linked blog post, (I still have i open in the next tab) and it seems that there was a lot of discussion (at least right after the renaming) that goes into problems new users (such as me and others who have posted apparently a lot of questions about). I think it was @kanga_who who kept mentioning “look at it from a new users perspective” and the fact that names were chosen that don’t really fit what/how it appears to anyone looking at using it and digging up copious amounts of YT videos and blog posts and forums on a self managed solution (old HA) to a complete OS/VM new style HA.
I saw quite a few references to “Home Assistant Complete” which is a good title if you want to think about it from an OS side, whether it be HassIO/Hasbian style on bare metal or in a VM, then Home Assistant Container becomes self explanatory, which just leaves “core”, again, pretty self explanatory.
I think the only mistake is having renamed a complete/OS/IO version to a name of something that was self installed on a generic setup. (I’m pretty sure I have it down now) /fingers-crossed
@ha-username: “Hass” (hate) is also known in parts of the English-speaking world. One of the first things I did (I’m currently living in Germany) is rename my machine to “has1”, i.e., “Home Assistant Server 1”.
Back on topic: Thanks @nickrout for bringing up The Last Word again. That at least brings some clarification.
So I want “Home Assistant Supervised” for all my machines and those of the lot of friends I have to support, right? Because it’d give us “the full Home Assistant experience on a regular Linux operating system.” Yay! Exactly what we want! Regular, in our case, being Raspbian Raspberry Pi OS (yet another name change, grr!).
That would bring Supervisor, Add-Ons, an easy experience for the newbie plus all the niceties of a real OS, like installing scripts, using system binaries via commandline sensors and automations, and installing/using, say PicoTTS?
So where to get and how to install that? Because “the Home Assistant Supervised method is here to stay”. And what about “missing for each of the above-listed are the specific details, e.g., the supported operating systems for a Supervised installation or the supported hardware”?
If you want to run an MQTT Broker like Eclipse Mosquitto, you have several ways to install it:
-
As an application
Go to mosquitto’s download page, get the version for your operating system, and proceed to install and configure it following Mosquitto’s instructions. It would be natural to use this version with Home Assistant Core, since both are applications installed directly on “bare metal” (so to speak). In other words, this is the traditional way to install apps (with all its inherent complexity). -
As a docker container
Go to docker hub and download the Mosquitto docker container. Proceed to install and configure it following Mosquitto’s docker instructions. It would be natural to use this version with Home Assistant Container because both are docker containers and designed to run in a docker environment. -
As an Add-On
Using Home Assistant or Home Assistant Supervised, in the main menu go to Supervisor > Add-on Store, click Mosquitto, and follow the displayed instructions to configure it. What you are actually doing is installing a Mosquitto docker container (like described in the second paragraph) except it’s a customized version and the installation and configuration process is much more convenient and transparent.
Yes, the new name is confusing. Yes, the project lead has agreed. Yes, a new (hopefully less confusing) name will be coming for the thing currently called Home Assistant.
@Tinkerer maybe OT in this thread but quick question… I followed your link and there no mention about virtualenv installs being supported. Do we need an ADR for virtualenv installs or are they just killing this completely now? I mean I understand the part about being a python app and they say it could never happen but I feel like it’s heading that direction. Already there are already a few “workarounds” required just to a have a clean install start without an invalid configuration error.
My thoughts : venv is not going to die. How would they develop on Home Assistant without ? A docker container is fine, but not for development.