Exec format error on fresh install of Home Assistant OS 16.2

[ahauser31]

Describe the issue you are experiencing

On booting into a fresh install (NVMe was wiped completely and cleared with all 0s before writing the HAOS image to it), after "Waiting for the Home Assistant CLI to be ready..." I get the following error:
/usr/bin/cli.sh: line 3: /usr/bin/ha: cannot execute binary file: Exec format error

The green "ha >" prompt shows up, but regardless of what command is entered I only get:
xargs: ha: Exec format error

Trying to access the web interface fails, even when trying to access it directly with the IP address; so I assume it is not starting

Switching to a different tty and logging in as root, I checked that I can reach the internet - ping, curl all work, IP address for the network card of the mini pc got assigned by my router, etc. All seems in order

Running systemctl status docker.service then shows what the issue is (probably):

ERROR (MainThread) [supervisor.docker.interface] Can't install ghcr.io/home-assistant/amd64-hassio-dns:2025.08.0: 500 Server Error for http+docker://localhost/v1.51/images/ghcr.io/home-assistant/amd64-hassio-dns:2025.08.0/json: Internal Server Error ("layer does not exist")
WARNING (MainThread) [supervisor.plugins.base] Error on installing dns plugin, retrying in 30sec

This then basically repeats every 30 seconds. I take it since it fails during installing the base system, nothing is working. Rebooting doesn't fix it.

What operating system image do you use?

generic-x86-64 (Generic UEFI capable x86-64 systems)

What version of Home Assistant Operating System is installed?

16.2

Did the problem occur after upgrading the Operating System?

No

Hardware details

GMKtec NUCBOX G3 Plus
Intel N150

Steps to reproduce the issue

1. Followed the generic generic-x86-64 installation guide using Ubuntu on a stick to restore the HAOS image to the NVMe in the mini pc
2. Rebooted pc, removed stick and waited for boot to complete
3. Get error message after waiting for cli to start
4. Change to a different tty, log in as root and check the status of the docker service
   ...

Anything in the Supervisor logs that might be useful for us?

Cannot access logs, as cli / web interface not working (or not starting)

Anything in the Host logs that might be useful for us?

Cannot access logs, as cli / web interface not working (or not starting)

System information

Cannot access logs, as cli / web interface not working (or not starting)

Additional information

No response

[Impact123]

Can you connect via keyboard and monitor / developer SSH or other ways to get a OS shell, enter login and try if this resolves it?
Since this is still a fresh install it might make sense to start completely fresh via ha os datadisk wipe once this works properly again.

[ahauser31]

Can you connect via keyboard and monitor / developer SSH or other ways to get a OS shell, enter login and try if this resolves it?
Since this is still a fresh install it might make sense to start completely fresh via ha os datadisk wipe once this works properly again.

Thank you for your reply. Just tried it but it unfortunately doesn't work... supervisior is getting stuck downloading amd64-hassio-dns with tag 2025.08.0 (messages in the output of systemctl status docker.service are the same as before).

Given that it's October already and supervisor is looking for an August-tagged version, was the dns plugin updated maybe, and supervisor is looking for an old version that is no longer available? That's what it looks like to me...

[ahauser31]

Any way I can install the dns plugin manually from the tty to help the supervisor past this error?

[sairon]

All the errors are indicators that the Docker data is corrupted. It's unclear what was the cause and it's pointless attempting to restore the integrity, as it appears that it affects files of multiple containers. At this point I'd go with a full reinstall of HAOS on the NVMe. Make sure the data is synced to the disk before unplugging it. If the errors reappear, check dmesg (or dmesg -l warn+) for any errors that could indicate issues with the NVMe drive itself.