Running in a VM. No changes were done, one day it worked and the day after i can’t access HA either from the android app or the browser. Restored two different backups, tried manually disabling an addon that was raising errors in the logs, no change.
Using the developer console in the browser i get
ns_connection_refused
when attempting to load api/websocket
I am using SSL with letsencrypt and duckdns, and the url resolves to my current ip without issues. Date and time are correct.
Addons appear to be running (i have frigate running) but i can;t access the front end at all and i can;t see much from the logs as i am limited to the vm display truncating 90% of the log contents and can’t scroll horizontally.
Of note, i get all the app notifications from my automations, but i still can’t access HA.
Although i have the core_ssh addon and that it appears to start (using addon start core_ssh) i can;t ssh into it – most probably because i restored a backup where it wasn;t set up yet), i am stuck with the extremely limited linux shell through the vm screen, without any way to set-up ssh or install any other linux package
How can i fix this ?
So i abandoned and used an actual folder copy of the vm i made a few weeks ago before updating, because it is probably the 6th or 7th time in a year I have to completely reset HA as it always finds a new and original way to break itself whether from an update or just because it felt like it. My last complete reset was on Jan 14. And no, my setup is not heavily customized. Just a couple automations, frigate and a couple addons.
Am now in the process of re-updating and re-configuring things, but mostly everything works other than being outdated. It would be nice to have a global “Update All” option instead on having to do multiple clicks and wait for every single item.
I will then make another physical copy of the vm folder as i can’t rely on HA’s backups, even when doing full backups.
Of note: this is running of a brand new ssd, but i somehow expect a filesystem corruption of the crashed install. Had some fsck warnings about conversion from ansi_something to another encoding, plus addons complaining about file errors. I had no power outages or anything, it just randomly stopped working.
Edit: After this, the app rejected my saved credentials and i had to re-add my phone and reconfigure my widgets. Hopefully my at home / away automations and alerts still work.
I’ve been using HA for two years. When it works, it is awesome. However it never works for more than a couple months without a major recovery intervention.