I added a NAS mount to my authorized directories in my configuration.yaml. This didn’t really help anything because HA won’t follow symlinks even to directories you mark as allowed. Anyway, I didn’t remove the directory from the config. My server did its weekly reboot last night and HA started in recovery mode because it couldn’t find the NAS directory. This is either an issue with the startup order or HA isn’t trigger the auto mounter. Seems like a bug that the directories in allowlist_external_dirs must exist for HA to start.
When I saw the issue, I removed the offending directory and restarted home assistant. It restarted fine, but every single integration, device and entity was removed. This was quite a shock.
Unfortunately, with HA core the built in backups don’t run automatically, and it’s been a while since I’ve created on of those. I have nightly server backups.
Why would recovery mode delete everything?
Is there anyway to restore everything without doing a system level backup?
Is is possible to make HA even less user friendly?
Use HAOS if you are worried about user friendliness. I would check the mounted folder for it contents, you might find the missing data there. If not restore your server.
I was able to successfully recover using a backup on my nas. Still, I think it’s a serious design flaw for it to delete everything following a “recovery mode”. I’ve been in an OS recovery mode countless times. When I reboot in normal mode I don’t find it’s deleted the file system.
I’d be in the same boat if I used the Container version and did the same edit to my config file.
I’ve HAOS, added trusted_users to config yaml, there was a problem with user_ids. HA went in safe mode, I fixed the issue, HA restarted “without problems” but now all integrations are gone. There’s no recent backup. Pain.
I’ve literally just had this happen to me - Had a typo in configuration.yaml, went into safe mode, entire configuration deleted with no backup - Is this is joke?! It’s going to take me weeks to get this back up and running again!!!
Another casualty here. Was doing the same thing as jmlt, adding the trusted users section following a guide to cast a dashboard to a google hub. Used the username instead of the user ID and now everything is gone. This is crazy.
Same here running 2024.5.4 in a VM Virtualbox.
I was working with auth_providers to enable an old ipad as a browser and due to some typo or wrong configuration, Home Assistant went into safe mode. After fixing the problem and re-starting all Integrations were gone!
Luckly I have a backup from yesterday and will try to restore.
It is really annoying. The Check Configuration button should do this type of check to avoid such incovenience.
Crazy. This just got me too. Same deal, I used the username instead of the user_id. Restart, recovery mode, fix problem, restart, everything gone. I belive I have a backup, but this is wild behavior.
Is there no recovery from this? Just deleted my whole system too, I thought I had setup backups, but apparently it was not running. Why would safe mode delete everything?
So glad you posted this @brianegge ! I too used the allow_external_dirs in my config.yaml and ended up in the same position as you. Faffed about for 2 hours trying to work out why my entire setup had disappeared and then discovered this post.
I totally agree - what the chuff is the point of recovery mode deleting everything - it’s beyond comprehension !
Thanks to a decent backup regime I’m back and running again but talk about unnecessary stress !
I am right now experiencing the same (running on HAOS) after restoring from a backup. The config contained the allow_external_dirs for a NAS, but the storage was not connected. So the system started in restore mode until I commented this setting. After restarting in normal mode, all integrations are gone…
Sure, there will be some workaround for this, but it took me quite a while to find out all this!
