HA has been installed on my Ubuntu machine inside docker for a long time. I just tried updating from 0.117.1 -> 0.118.4 and my supervisor is telling me:
Here’s my docker list in case that helps ID what I’m actually running, because I am confused.
EDIT: I followed this as my issue seems to be the same. It’s merely a stop-gap and will probably have consequences in the future. Can someone please explain what I’m supposed to do to fix this completely?
This appears to about to be a big problem for lots of folks. I too am running in a Docker container on top of Ubuntu 20.04. It’s been extremely stable…
What’s even odder is that I’m running HA 0.118.4 and just trying to upgrade to 0.118.5, but suddenly it won’t let me. Looking at the Supervisor I too now have “Your installtion [sic] is running in an unhealthy state”. What this appears to mean is that the Supervisor is not running with privilege.
There are commands here you can supposedly run to correct this, but I’m wary of doing it on a working installation. Elsewhere in the community folks are saying a complete re-install. Like we have a weekend to kill.
I’d like to understand why this has “suddenly” occurred and why the change was necessary?
To be blunt Supervised is no longer supported on Ubuntu as was telegraphed many months ago it may stop working at some point. It seems the latest version of supervisor triggered this. The solution is to switch to Debian which is supported. As a stopgap it seems you can fix it with the installation of a package mentioned in this thread Unhealthy state
You can complete a Debian 10 install plus HA Supervised, restored snapshot and be back up and running in under 1 hr. It doesn’t take a weekend.
That is because of a Supervisor update, it’s not to do with the version of HA. The Supervisor automatically updates, and this check and output of error is obviously deliberate with this update.
HA 1.0 is due shortly, so it makes sense this change was made now to force anyone with an unsupported system to move a supported system before 1.0.
I was faced with the same situation as you. I upgraded to Ubuntu 20.04 and that eliminated the ‘unhealthy’ indicator. The upgrade introduced a more recent version of Network Manager which is important because the one in Ubuntu 18.04 is 1.10 which is lower than the minimum required by Home Assistant (1.14).
The upgrade took about 90 minutes (will be much faster if you have an SSD). However, the long-term solution is to change from Ubuntu to Debian because that’s the only officially supported Linux distro. I plan to do that after I purchase an SSD to replace the existing hard-drive.
Yeahhhhh HA is not the only thing running on this server. This is actually motivating me more to buying a second server just to isolate HA away from Plex and other things. Not a big fan of being told to install a new entire operating system to get support maintained.
That part takes more time, obviously, but hopefully you have kept notes and text documents with your own installation instructions, docker-compose info you can copy/paste, etc. Makes life very easy and reduces complete setup time significantly. I run Plex, Shinobi, mqtt server, Transmission and a bunch of other stuff. I can set up a new machine with everything I run, shared drives between VMs, VPN etc in under 2 hours, copy paste and done. I know that’s a bit OCD, but I like to keep notes for everything and how I got it to work.
Not a silly idea, might be worth getting a Pi4 if that’s all you will run on it.
Unfortunately, we don’t get to decide the conditions of how HA runs, the Devs do. The information has been out in the wild for a long time letting everyone know this could happen at any given moment. The moment has come, time to change or stay on 0.118.4 indefinitely. 
Yikes. I’m sure lots of folks are running HA on a virtual machine with Ubuntu (as that’s how it is taught by Dr. ZZZs and The Hook Up youtube channels).
FWIW, I ran into this today as well, Ubuntu 20.04.1 with a lot more than hassio on my box, in the end I had to remove watchtower (both the running container, and the image of the filesystem). After that, the upgrade works again. So it still works on Ubuntu 20.04.1 for those wondering (for now).
This worked for me as well, I deleted the container and the image, then rebooted. The unhealthy state went away after.
