Migration/re-install of system (from hass.io in Docker on Debian to hass.io in LXC on Proxmox) - possible caveats?

Hi all,
due to overall degradation of my system, I’m planning a re-install of things…

While trying to figure out how to get this done the easiest way, I’ve been thinking about the possibilities and/or risks using partial snapshots for some of it and hope the community can help me out :slight_smile:

TL;DR
Is it possible to “migrate” well-functioning add-ons (Unifi Controller, Bitwarden RS and Nginx Proxy Manager) from a Docker-based install to a fresh Proxmox VE-based install of hass.io using partial snapshots?

Present system/setup
Intel NUC NUC6CAYH (Celeron J3455) with 120 GB SSD
Debian Stretch
Docker-based hass.io including various add-ons

Three hass.io add-ons are playing a major role in my entire setup (UniFi Controller, Nginx Proxy Manager and BitWarden RS) and are all behaving very well at the moment.

Planned system/setup
Same hardware, with the addition of an extra (external) SSD (location for databases and backups)
Proxmox VE
LXC-based install of hass.io (using whiskerz’ script).
Debian+Docker+Hass.io as described by Tomas Hellström (@helto4real) here

Intentions
I do not want to keep my present Home Assistant as a whole, but want to start over with a new install (including a complete reset of z-wave and zigbee).

I’m aware of the procedures described for migrating from an existing UniFi Controller at the Community Hass.io Add-ons: UniFi Controller as well as well as the instructions on backing up your vault with Bitwarden.

Challenges/questions
My question is, if it is in fact not only possible, but also far easier to use (three) partial snapshots of the add-ons in question - instead of going through all the export-import-stuff…
However, as there’s a lot going on with certificates and keys in the three containers, my guess is, that a lot of things can cause trouble - especially when changing from one OS to another…

Unifi Controller
The procedure for migrating the Unifi Controller includes having access to both the old instance and the new instance at the same time. That will not be possible in this case, since I format the SDD during the proces. I will of course make a backup from the controller software that can be used, but I guess there is a risk that the Unifi Security Gateway (and other devices) will not “accept” being provisioned in the new system making it necessary to factory reset them all in order to provision them in the new system (not a big deal, only time consuming)…

Bitwarden RS
The native backup feature in Bitwarden doesn’t include a “complete” backup of both organisations and all users, but only “per user” and “per organisation” and the result is valid, but turns up somewhat messy and includes users having to validate again. Yes, of course I will make backups, as loosing the vaults would not be good, but still…

Yes, I could just go ahead and try - but my hope is, that planning ahead (for once) can prevent me from wasting a lot of time…

So, any inputs anyone?

You might want to check a recent blog post about that. That script is not recommended by the devs here.

I’ve just used that script myself as an experiment and I’m going to instead install Debian+Docker+Hass.io

Thanks for the heads up on that, @DavidFW1960!
Will Proxomox VE -> Debian -> Docker -> Hass.io - better safe than sorry…

Not sure actually… I suspect it would be. I setup a Debian VM in Proxmox yesterday but it’s doing some weird things when I ssh in… I don’t have anything even installed there yet but the ssh session keeps bailing out… I might have screwed up some permissions or the firewall maybe… dunno. Need to do some more reading.

A brief upodate in case anyone could find that useful:

Proxmox VE 6.1-5 :arrow_right: Debian 10 :arrow_right: Docker CE :arrow_right: Hass.io 0.103.2

Restoring Bitwarden RS addon and Nginx Proxy Manager addon from partial snapshots worked like a charm (think including the “SSL folder” and “Share” folder is wise).

Also restoring “UniFi Controller” went well - however, an important detail is, that when restoring from the snapshot, it seems to expect that the folder “/hassio/backup” includes the subfolder “unifi” that includes different files (meta.json, e.g.) - and without that folder being present, restoring was not possible.

In the section Automated backups in the description of the addon on GitHub, it says:

“The UniFi Controller ships with an automated backup feature. This feature works but has been adjusted to put the created backups in a different location.
Backups are created in /backup/unifi . You can access this folder using the normal Hass.io methods (e.g., using Samba, Terminal, SSH).”’

Anyway, the grand total is, that it worked out pretty well :slight_smile: