I need to vent about what has become the most frustrating experience in my years of using Home Assistant: the backup and restore system is unreliable to the point of being dangerous for anyone relying on it for disaster recovery.
If you have a small, basic setup, you’re probably fine. But if you’ve built out a real smart home with integrations, automations, and customizations? Good luck.
I’ve been trying to restore a legitimate, valid backup for over 24 hours now. Not on some janky SD card setup - I’ve tried:
Raspberry Pi with SSD attached
A dedicated NUC server running Home Assistant OS
Both setups? Stuck on “Cleaning Home Assistant Application” indefinitely. Hours and hours with zero progress past that screen. The backups aren’t corrupted - they’re just large. Apparently that’s too much to handle.
This is a Critical Failure
Here’s the thing: backups are not a nice-to-have feature. They’re fundamental infrastructure. When your backup system fails, everything else is built on quicksand.
I’ve run out of ideas. I’ve followed every guide, tried multiple hardware configurations, waited patiently for processes to complete. The backups are valid. The hardware is more than capable. But the restore process simply cannot handle anything beyond a basic setup.
I can’t even manage to get an unencrypted version of these backups despite knowing my password.
That said you don’t say anything about what you actually tried. So it’s kinda hard to help.
I personally have a rather uh chonky install. And my backup is 14G without media and I do a DR exercise quarterly. Besides having to turn off encryption once it’s been relatively painless.
Yes I have seen my restore appear to hang (poor UI on the restore) but after leaving it and rebooting that instance was essentially it never came back and said it was done.
I don’t generally load the backup during the onboarding I just do a vanilla install with a generic ID of latest, login. And restore from inside the backup facility rather than onboarding. It seems to work better.
I get your frustration, but there is plenty of evidence saying otherwise. If anything, online guides from any other source than the offical docs tend to be outdated.
Even very big restores tend to go pretty quick and seamless. Take a deep breath, and follow Nathan’s advice. We’re here to help.
Thanks, I appreciate the assistance and offer to help. @NathanCu@Edwin_D
My backup is large 28GB+. I’ve tried restoring this on both the original rasberry pi with SSD attached which it came from as well as a VM running home assistant OS on a dedicated NUC server. Both land stuck on the below screen.
So now I have the unencrypted backup, I feel much better that worst case I can rebuild from the config files.
I don’t even need the historical data so what I’m trying to do now is to recreate the backup tar file without the DB included in hopes this restores much quicker.
Happy to provide additional details and logging, please let me know where you think I should go next.
Edit: I failed to mention that the restore is being done post clean install via system → backup → restore. I am not running this during onboarding.
Were you restoring into a clean install or over an existing config? I had a similarly sized one. As Nathan said, From the onboarding screen I too had backups seem to hang. Turns out it wasn’t doing anything and not telling me. I also did a clean install, login, use Samba to copy the backup,to the official backup location and restore from there.
Another thing: could it be you are using Adguard or Pihole or similar that is blocking the setup from fetching things?
If you can access the Home Assistant CLI while it seems to hang (e.g. by attaching a keyboard and monitor, or ssh) in the CLI try typing: jobs info
It gives you running jobs. If you see an empty list, it is not doing anything. If it is busy, you see what jobs it finished and which are running (such as restore parts of HA or addons)
ps. I’m about to do some training, so I may not respond for an hour or so).
@Edwin_D no worries, thanks! Yes i can access cli via keyboard and monitor. I have run the jobs info command and when I thought it was hung I did see jobs. Is there any way to get more details on where those jobs are? How long should those jobs take?
A couple hours on my longest one never waited more than 3.
(Also worst case if you have the backup decrypted it’s simply a copy everything in the config folder including the hidden folders. Expand tarball + copypasta )
Were you using the same SSD on all attempts and are you sure it is ok, using a fast data cable (if it is USB) and sufficiently powered? If you picked the wrong cable USB can be sloooooooow.
That IS your config folder. (see configuration.yaml) Show hidden there should be hidden folders beginning with .
You copy that and it’s ENTIRE contents including any hidden folder over the top of the config folder and restart. If you don’t want the dB don’t copy the dB.
With regard to copy and “restart”: It may or may not be obvious, but copy it while HA is not running. For instance using Samba addon. So stop, copy (including hidden files/folders) and restart.
However, I’m not sure this way all addons you had will also be installed. So if the restore looks k but just slow for whatever reason (as in you see changes after you ran job infor a couple of times) backup / resture still is the safest. HA resture itself was very quick for me, the addons took a good while longer. But I could see the job list shrinking, so I knew it was progressing).
Yes, plus you have add-ons config, so add-ons should install effortlessly. IIRC addons config is a separate share on samba, I’m not sure if that links to the config folder. It may need copying elsewhere if a partial resture of addons does not work as expected.