I am currently using HA dockered and I am considering moving to HA OS to simplify my life. I would like to have HTTPS setup to access my HA OS.
Here is my current setup. I have Nginx Proxy Manger setup in Docker on my server. Now I could spin up a new VM and install HA that way or I have a Tiny PC with an i5 collecting dust that I could use. Either way I am not sure how to setup NPM to grab the cert from Let’s Encrpt and be a reverse proxy for something outside of a connected docker container. I don’t even know what to search for to find out more information on how to achieve this setup.
This is really the only thing holding me back, has anyone done this or have a good resource I can reference?
Do you know if that plays well with other NPMs? I thought about using that plugin but I don’t know how I’d handle the redirection for the sub-domain to the “external” server.
I guess I don’t know why moving to HassOS would simplify your life?
you still have just as many moving parts since you still have docker at it’s core and you still need to install and manage all of the add-ons just like you would have to do installing and managing the equivalent docker containers.
Especially since you already have a fully functioning dockerized system. All of the hard work that HassOS purports to make easier (debatable) is already done.
The only real benefit are snapshots but that can be replicated. I personally just copy my config directory after every big code change or right before an update.
And not to mention that you can still have all of the benefits of HassOS by running a Supervised version.
If HassOS actually was so much better than the other alternatives then I would think that all of the developers would be using it as well but from my understanding they mostly use docker as you and I are.
I am just wondering what you think are the benefits?
There are two docker containers that might be of use to you;
linuxserver/swag
linuxserver/duckdns
I also did the move to HA OS because I’d bought a HA Blue and so moved it off of Ubuntu/docker setup to its own hardware. I installed the linuxserver/swag and linuxserver/duckdns containers that ensures that I get an updated duckdns certificate from letsencrypt and configuration for the nginx setup. There is good documentation supporting.
That leaves my ubuntu/docker hardware to run a plex etc etc setup without impacting the HA setup. HA runs clean as a nut since then.
So my drive on this is that for the past year I have had very little time to put towards keeping HA updated. I work in Healthcare IT, by the time I get home I have little interest in doing a lot of config or maintenance on anything. I do like to work in HA though, I am able to impliment things that make things easier and bring happiness (I’m looking at you Holiday Lights) In fact I am still on 116.3 because I’m dreading what’s changed and what’s going to break.
So new install on a new server, while there will be an investment of time I can keep the current working HA up while I get the newer one up and running and then do the switch over.
I looked at HA supervised by I don’t think I want to invest that much effort in maintaining something that seems so fragile that if I apply a docker update by mistake or as part of a security patch my HA setup will fail.
I like the idea of a simple click install for Add-ons, but I’m wondering if there is any advantage to Add-ons and maybe I’m not fully understanding what Add-ons is. I thought it was something closer to HACS but it appears that it just lets you install things like Plex, MQTT, Node-Red, etc… Things I already have installed in docker.
I really thought there would be less maintenance. Supervisor would keep my install up to date and Add-ons would make adding stuff to HA simple and prevent me from filling my configs with bits that I missed when removing components that stopped working and/or were abandoned when a HA update was released.
Switching to HassOS won’t help you with that at all. zero. nada. You still have to do all of the updates (if you want them) and you will still have to deal with breaking changes just exactly like you do now.
If you have another machine to use to do the switch over then it doesn’t make any difference if you run HassOS, HA Supervised or HA Container. The end result will be the same. The only difference will be the final HA install flavor.
I have a test install of HA Supervised alongside my production HA Container install (and another HA Container install that I use for other testing - and all on the same NUC machine) and it has never been affected by any system updates. Of course, I can’t say that will always be the case in the future but we can’t even say that about HassOS either since there have been a few times when an automatic Supervisor/OS update left people’s HA installs completely broken in the middle of the night.
And to put a finer edge on my statement above - as far as I know the HA Container install is the easiest way (if not the only way possible) to have multiple installs of HA to test things on if you don’t want it to impact your production system.
So since you are running in a Docker environment you could load up another HA Container instance on your existing machine, do all the updates to get it current and then switch to the new updated system for your production system without needing another machine to run it on during the switch. Again, I’m not even sure that’s possible with any other install method. But it’s definitely by far the easiest way.
Yeah, that’s literally all add-ons are - Docker containers of different apps that are set up to be managed by the supervisor.
But “managed by the supervisor” isn’t as easy as a “click to install”. There is still some configuration required for probably most add-ons. And in the end you still end up with those apps in a docker container just as you have now. HACS is completely different in that it adds stuff directly to HA itself just as any other integration or lovelace card. add-ons =/= integrations at all.
And you still need to update those add-ons as updates get released. There is an auto update option you can use but I personally really dislike the idea of auto-updates of anything because there is a history of some auto-updates breaking systems.
I can’t say that add-ons don’t make things a bit easier to set up and manage the resulting docker containers (but only for add-ons) but if you already run Portainer then you’ve already got 90% of the battle taken care of.
The Supervisor will only ever keep the Supervisor up to date (unless you opt for the auto-update for add-ons too). And YOU CAN’T OPT OUT of that auto-update for the supervisor so you get the potential downsides of a possible broken supervisor auto-update.
The OS will also (I believe) get auto-updated as well (which I’m pretty sure is not optional either).
HA itself will still need to be updated manually on every release just as it is in your current install.
That has nothing at all to do with OS/Supervisor.
You will have the exact same issue using any HA install method.
Seriously thank you breaking down what ever does and doesn’t do. I’m pretty sure I would have been seriously disappointed has I gone to HA OS. I really did get the impression that a lot of things were automatic with recovery if things went south during the ‘update’.
As for a seperate machine or leave it where it is and create a new container, you’re bursting that bubble too
So finity what do I do with an i5 Tiny with 8GB of ram now?
Core in a venv. You can have as many venv’s as you want with HA in every possible flavor installed side by side. They’re just directories. It really doesn’t get any easier than that.
In fact I am still on 116.3 because I’m dreading what’s changed and what’s going to break.
