Help! Error in configuration.yaml and cannot access Home Assistant Green

Installation
1

I’m using a Home Assistant Green. After editing my configuration.yaml to add an external_url entry, Home Assistant failed to restart. On reboot, the device only shows a > prompt and doesn’t fully boot — the web interface, SSH, and add-ons are all unavailable. I can’t even run ‘cd’ at the prompt. I believe the YAML config change broke startup.

I cannot access the internal storage to fix the file, and I understand the Green does not offer UART or other direct access. I don’t have a recent backup.

What are my options? Is there any way to recover without losing everything?

2

You are probably at the HA CLI. Execute this prompt to drop to the shell:

login
3

Thanks, but I don’t think I’m at the HA CLI - the prompt is “>”, not “ha >”. If I type “login” it doesn’t return anything - not even another prompt.

4

Do you have a backup from before the change that you can restore?

5

No unfortunately not. Obviously regretting that now! However, my configuration.yaml is very simple - so if I can access it I’m sure I can fix it.

6

When you reboot HA it checks your configuration and will let you know if it’s invalid. With that said, just because it says it’s fine doesn’t mean there isn’t issues. I think a lot is checking formatting, especially for YAML but no matter what you did to your configuration.yaml file I don’t think it would break HA. It may break the front end (GUI), I’ve done that more then a few times but could always access it via SMB or SSH using HAOS with the ads ons for Samba share and Advanced SSH. I could always access the files over the network.

If you are using the m.2 slot it would be worth scanning the drive. I just don’t think there is a bootable Ubuntu or Linux USB drive you can create on that specific hardware. Maybe there is a generic ARM version out there. Other option would be if you had an adapter to scan the drive or the BIOS has some sort of disk scanning.

No matter what you did to the configuration.yaml, it shouldn’t cause the behavior you are seeing. It may break the front end but there is a lot more to HA doing a clean boot then just loading what’s in the user directory (config, SSL, share, ECT…). That’s why restarting HA is typically pretty fast, you aren’t doing a full machine reboot. Just HA Core and supervisor I believe (correct me if I’m mistaken ) with HAOS. I do know it’s not doing a full reboot though. If it has some sort of BIOS at least see if you can see the drive unless it’s using internal storage. I’m not that familiar with the hardware outside it being AarM based. If it doesn’t get to a point were you can ping it something else is up than anything done to your config.