Docker vs running natively on OS

I am still very new to Home Assistant and first started out installing it on Debian in the “env” install method and I loved it. It picked up so much stuff in my local network it was amazing.

Playing with it for a few weeks I found that I wanted to do something, but I needed to have the Supervisor tab which I didn’t have.

That meant I went and installed the VM image and got it running however it did not pick up anywhere near as much as my previous install. This is when I first heard of a thing called Docker.

I get what Dockers are, but find that it seems to add another network layer between my network and the Container, e.g. looking at my VM console I can see that the Docker Container has a 172.x.x.x address and the network address I have given it on my network, 192.x.x.x

This also added complexity when I was trying to find the configuration.yaml location, because I didn’t understand (I do now) that at the console you are in the container, but when you SSH in you are in the OS that is running the Docker.

Reading through the new post about installation methods supported vs community it would appear that all the new supported methods are Docker only.

Leads me to a couple of questions, have I miss understood how the Docker works and why is there not a full (supported) OS installation that can be installed natively?



There is, it is called home assistant core.

1 Like

Yes, but Home Assistant Core is missing Supervisor and Add-ons, so what is the point really?

This would appear to be the version I had installed initially, and this is where my confusion was when I first started as to do something, I don’t recall exactly what I was trying to do at the time, but it REQUIRED Supervisor. So I thought I had a fully supported version but I did not.

Also you have to do this version yourself, you have to install Linux, get PHP installed and then install HA, that is not what I am asking about. I am talking something that has the OS included with all the requisite packages, download image and run install away you go.

Then you want Home assistant OS.

I read all your posts in this topic and it seems clear to me that you made several incorrect assumptions about the installation methods. Did you read the documentation’s section on how to install Home Assistant and was it responsible for the confusion that it has caused you?

If the answer is yes, what improvements would you suggest to help prevent other users from going through the same frustrating experience?

1 Like

Its mostly about dependencies. Addons are not plugins or configurationof home assistant but mostly other applications etc with their own dependencies. And if you start installing a lot of software on linux with dependencies it is easy to fall into “dependency hell” where one dependency prevents another from working. Docker helps with this at the cost of yes, abstracting things like network, storage, etc. I just think of it as docker is the packaging method for addons.

You could do this yourself on the base os but you lose the packaging method. There is nothing preventing you from installing Samba or Node Red or whatever and just configuring it manually. I did this for a while. I went from straight OS to a Python VE, but on upgrades it all got so unbearable I would usually end up having to rebuilt most of it.

I personally use HA on Docker, I live just fine without addons or a supervisor. I have Node Red, Samba, Portainer, Mosquitto and a bunch of other containers installed from docker hub. I configure and maintain them individually, HA doesn’t know the difference. I back my stuff up, I control what versions get installed and when and on the downside I usually invest a couple of weeks in getting HA working again if and when I decide to upgrade it. It is more work but I prefer to have more control over things than the full HA Supervised method.


People, thanks for the replies.

@123 I will be up front and honest and didn’t really read much of the installation stuff, if I did I don’t remember it. I am more of a kinetic learner and because I had my own preconceived ideas on how I wanted to do things I went the way I was most comfortable with.

With the Docker container I think there is some documentation missing in that I have some integrations that you require to have to add them to the configuration.yaml manually but I cannot find how to access this file in the Docker instance. In another thread it is suggested to turn on SAMBA, but to me that just adds overhead that I would never use.

Looking at accessing the Docker via the console even that confuses me a little, warning I can be very literal, it shows the usage as “ha core [command]”, but when you run something like “ha core info” it doesn’t work it just show you the help again. you do “ha info” it says it is depreciated and use “core” instead. So when you do “core info” you get the same information, yet it confuses me the help info.

So I freely admit I didn’t read the all the documentation but I don’t honestly think it would have helped me understand the difference in the installation methods as my thinking is irrespective of how the software is installed how you access and configured it should be the same.

One experience no matter how you get there.

It’s pretty much impossible to have more control than HA Supervised method…

Like I said in my first post, you made several incorrect assumptions about the installation methods. Now we know why that happened.

Your experiences with Home Assistant will continue to be a source of frustration if you avoid reading the documentation and rely exclusively on your erroneous assumptions about the product. This attempt to just “wing it” has not served you well. I encourage you to review the section on installation methods.

Good luck.


So what you are saying is that depending on which way you install the software will depend on which way you configure it.

On reflection I did read the installation documentation here: downloading the VMDK and I was able to get the software installed and up and running. I didn’t wing it, the documentation is very light on and it doesn’t explain any of the questions that I have asked.

I appreciate that it is frustrating that you take the time to write all the documentation and people don’t read it, but sometimes people just need to be pointed in the right direction.

Trying to understand why HA is working a certain way and as I pointed out even the documentation in the console is not consistent, makes it hard for someone with some semblance of IT skill to have a play with something.

Also I read through this documentation: and find that the configuration directory is in different places depending on your install. This document looks like it has been updated since I last looked at it, but again it didn’t clarify my earlier questions.