I have been running hass on an ubuntu server for over two years and have been using the upgrade features of the core, the supervisor and the add-ons over this whole time.
Recently, that stopped working. The system is flagged as unhealthy and the log seems to interrupt the upgrade process.
Do I have to re-image my whole system on debian and reinstall everything from scratch? How can I for sure debug what is causing the system to be flagged as unhealthy?
Ubuntu is unsupported since May last year.
Personally I wouldn’t bother trying to keep it working with Ubuntu, it will always be workarounds and maybe at one point you won’t even get it running anymore. Bite the bullet and change to Debian.
No, make a snapshot of your current system, install debian, restore snapshot.
Normally in the supervisor tab you can click on the link next to the message that your system is unhealthy and it should tell you why.
Ah, okay, What I do find odd, though, is that this has not been a problem for the last two years and I’ve been able to upgrade. Is this something that has recently changed?
Take a look at this. It may not apply, but is worth a try.
As I said the change to make it unsupported was May last year, lately more people reported issues with unhealthy state and not being able to update, this was to be expected.
Yeah, I guess I missed that memo. I did not notice that Ubuntu was listed as unsupported when I first set myself up with hass. If Ubuntu had some problems running hass, I never experienced it. What I am experiencing now is code that was added that prevents me from using it as I was before.
I have been running home assistant on a big server that runs a bunch of other services and switching the whole thing to debian would be a major undertaking. It would be nice to have the ability to continue using my installation at my own risk rather than being dead in the water.
Additionally, I am also no longer able to restore old backups for the same reason. In the process of trying to get this to work, my addons are no longer being loaded in HA core-0.118.3 which is the one that was running before this bug showed up, but since I can not restore the old setup, my home assistant instance is non functioning. I am really stuck here.
This thread is basically a copy of this I feel screwed over: Ubuntu suddenly dropped, All github issues being closed and im stuck with a broken setup and no way to fx
Read through it and you’ll fixing your ubuntu install is not recommended by anyone.
A big server only running one OS, no VMs or LXC Containers? You could make a backup of the entire system, install Proxmox on the server and run it as a VM. Then backup Home Assistant, spin up a separate VM for Home Assistant with a supported install like Debian or the full image with HassOS, restore the HA backup and never bother again with unhealthy or unsupported stuff. And you’ll learn some interesting stuff on this journey as well.
Sorry, no, I was unclear. I meant to say I have a single big server running a whole bunch of docker containers. I did put log a ticket on github which Franck closed but he did give me some hints that eventually allowed me to narrow in on the problem and get it working again. Two of the non-home assistant containers I was running were causing problems and I think I had some permissions issues which were resolved when I re-ran the supervisor installer script. It seems to be operational again and I now know that I need to re-image my host with Debian but I at least have a working installation for now.
I did take a look at Proxmox and will study it further. It might be a bit above my pay grade but as you say, I would definitely learn a lot.
Thanks for the tips.