What is the difference between Hass.io & a docker install

[txNgineer]

I am currently running Hasbian but have Docker installed and use it for other apps. What would the difference be for me to go to Hass.io and load other docker apps on it versus just putting HomeAssistant on a Docker instance over Ubuntu? I would really benefit from seeing a brief pros and cons list.

[lolouk44]

hass.io is designed for Raspberry Pies, but otherwise it’s pretty much the same. hass.io runs off a docker container.
Main differences are that hass.io has an extra container to manager the main one (i.e. taking care of updates etc) and hass.io has plugins.
I personally have hassbian on a docker on a Ubuntu box as it means no need to maintain required 3rd party apps.

[bachoo786]

Hi @lolouk44

I am fairly new to docker and have ordered a nuc and will be running ubuntu on it. I wanted to find out the following:

Is it possible to share

1). how you installed HA on your docker?
2). how you upgrade HA on your docker?
3). is the config folders setup? if no how did you do it?
4). what about custom components?
5). have you got node red or appdaemon running on your linux? if yes how did you link it to HA?

Thanks.

[flamingm0e]

@bachoo786

how you installed HA on your docker?

It’s a docker image. You just run the command from the home assistant docker installation page.

@bachoo786

how you upgrade HA on your docker?

You just delete your image and run the command again

@bachoo786

is the config folders setup? if no how did you do it?

??
You have to have a config directory somewhere, and it is on the host in a mapped volume.

@bachoo786

what about custom components?

What about them? You create a folder inside your config directory called ‘custom_components’ and drop them in.

@bachoo786

have you got node red or appdaemon running on your linux? if yes how did you link it to HA?

Node-red will run just fine in its own docker container, and you link it using the node-red home assistant node…it just uses the IP and a token or api password you set.

[lolouk44]

Install docker-compose
Then I use this docker-compose file:
https://github.com/lolouk44/homeassistant/blob/master/docker-compose.yml
Note I’m also using Grafana and InfluxDB. Just ignore these if you don’t want them. If you want them,this is a great guide:

Complete guide on setting up Grafana/InfluxDB with Home assistant using official Docker images

I know there are already a few tutorials on setting up InfluxDB and Grafana with Home Assistant, but they did not meet my requirements. Specifically, I did not want to use third-party images, which may not be maintained, and not use panel iframes to display the plots. Moreover, many of the tutorials still show the InfluxDB admin GUI, which is no longer available. So, here’s my version of how to get everything set up with the official images. Run the following command to start the InfluxDB con…

then to start HA I simply write: docker-compose up -d in the terminal from the folder docker-compose.yml is located
you may want to use something like portainer to manager your docker containers.
Check this post that has loads of good resources:

My Docker Stack

A lot of folks ask how to get started on a Docker stack and what to use to manage it all, so I figured I could share my docker-compose file for everyone to dig through and modify to your liking. There are plenty of writeups on how to get Docker running on your platform, but I recommend using the official documentation from the Docker website. I assume most of you might use Ubuntu, so here are the instructions for Ubuntu A few notes about it: On my docker host (Intel NUC, 8GB RAM, 256GB Sams…

[txNgineer]

Your approach was the direction I was leaning toward. Currently I have raspian on the Pi and have loaded Hass, NodeRed, Mosquitto, and other apps on it directly. The docker approach makes sense because it will be cleaner to maintain and update. I just wanted to make sure I was not missing something else from a hass.io build.
A side note: I looked into the differences between Docker, Vagrant, VirtualBox, and FreeNAS Jails (where my data lives). It looks like the Jails have the advantage of clean memory and resource management along with the hugely robust ZFS system. VirtualBox is fine where I have a bare metal server that I want to virtualize and plan to manage resources individually. Vagrant looks to be useful as a more flexible alternative to VirtualBox if you are only wanting to spin up and remove various virtual servers. Docker is a very clean way to sandbox all the various apps that we need on one server. For me this means my fixed services like PFSense and Bind+Postfix will be in separate VirtualBox VMs on a commercial class server. Then my data storage will be in a FreeNAS Jail and will hold MariaDB, Influxdb, Grafana, and Mediawiki. Finally my automation and IoT things will link to the RPi3 which I will convert to a Docker install and have containers for Portainer.io, Mosquitto (MQTT broker), NodeRed, and weewx, etc.

Thanks to everyone for the input.

[flamingm0e]

@txNgineer

For me this means my fixed services like PFSense and Bind+Postfix will be in separate VirtualBox VMs on a commercial class server.

I would suggest you run a proper hypervisor. Virtual box is not what you want

[txNgineer]

Been using VBox since its Innotek days in 2008, so just never had a need to change. I am perfectly agreeable to a better hypervisor if it meets my key requirement: A) is open-source (GPL2 or equiv), B) is free (as in beer), and C) has good, clear documentation.
What would you suggest?

[flamingm0e]

@txNgineer

is open-source

Virtual Box isn’t open source. So that’s a requirement?

I would say KVM on Linux, like Proxmox

[txNgineer]

Thanks for the tip regarding Proxmox. It looks promising and I will start with their videos and documentation.

Perhaps I misspoke regarding “opensource”. I know that it started that way and seemed to still be licensed as such by Oracle. On their web page it says, " 1. How is VirtualBox licensed?

The VirtualBox base package contains the full VirtualBox source code and platform binaries and is licensed under the GNU General Public License, version 2. You can distribute and modify the base package, provided that you distribute all modifications under the GPLv2 as well.

I am not a strict open-source person, but I do want the core of what I use to be licensed GPL2 and in a position to minimize the chance that a big corporation will flip the terms and cut the community users out. For example, like IBM just did with WeatherUnderground.

My biggest concern is that I am able to personally be comfortable with the level of privacy and freedom provided by things I put on my machines and network.

Once again, thanks for the info on Proxmox.