I have a backup from before and it does seem to be working, but every browser and the App sits “Loading Data”…
What Now? Months of building and configuring gone because of a system update?
I can SSH to it and that’s very restrictive being the HAOS it’s a virtual environment.
It’s still turning on and off lights etc, so it’s the GUI interface that’s failing, plus a loss of all the feedback I give it lost since 10AM this morning…
This happened a few years ago, I still haven’t got Knightsbridge devices back yet. Do I need to rebuild again, just to restore a backup, a backup I can’t get off the PI4 because it won’t auto mount a USB stick, and the NAS is read-only!
The Observer URL works on 4357, so what have they done with 8123 and the entire frontend in this last update?
I have done two sudo reboots which will restart everything, including the core.
SSH proves networking is up along with the operating system core and the lights turning on and off tells me it’s running. It’s the frontend that gone.
I will try as you suggest…
Update: ha core restart
Response: -bash: ha command not found
I thought it used zsh not bash?
Update: managed to scp tarball backups to another computer.
Please tell me I don’t have to rebuild just to restore from a backup?
What have they broken, it gives a login screen, so it’s running the Web service. There’s some changed to protocol, action, confirmation or data retrieval that’s foobar or a broken link, What did they do in this latest update, it must be trackable if they have source control…
Is anyone else suffering this issue or is there something unique with my HAOS implementation?
it’s a very specific failure for a hardware error, besides it’s an 200GB SSD via USB.
It’s the update that’s caused it, it was fine before the update and lost connectivity via web services after.
Are you running HAOS 14.2
Core 2025.2.2 (the offending update)
It’s up running, turning on and off equipment. Provides web services to login but then hangs “Loading Data” that’s not hardware it’s software and somebody has screwed up. They change a setting it list a network route in the virtual crossover. Something… But it definitely came from Core 2025.2.2 being applied.
I’ve got no interface so I can’t turn on kitchen equipment with no physical switch and I can’t log certain things like BP and health stuff with no interface.
I’ve recovered the backup from just before and 5AM this morning. I hate to think I have to rebuild the entire system just to recover the backup. It kinda defeats the object of a backup. Plus I can’t trust the update again, it’s a quandary. It’ll take hours to rebuild and restore, only to have it happen again?
The 1.3GB database should be intact but I lost everything from 10AM until it’s rebuilt and functional, and I’m in no state to do that now.
I’m extremely angry that their update did this and there’s no recovery option without the GUI they’ve denied me…
It the second time this has happened, last time it shed devices over a period of a year until it was effective non-functional. Now it’s full working but inaccessible…
Testing is done by the beta team of users. I am curious how many people are in the beta group, but they hang out on Discord which I will not use, else I would be one of them.
The fact is you seem to be the only one with this particular issue.
If an update bricked your installation, reinstall Home Assistant as if it is new, then restore from your most recent backup.
I’m becoming to dependent on such an unstable system, this is not the first time I’ve had to rebuild for something caused by them. If you can’t recover from a backup without a complete reinstall there’s something wrong with the backup.
Is there no recovery without a GUI?
Without solving the issue it effectively means upgrading in step with their release is impossible because it will just brick my entire system again.
To be clear it isn’t bricked, they’ve changed some protocol, firewall, routing between virtual and physical or simply lost the data feed. It is working, it’s just inaccessible having no GUI. It logs in, so it is running the service. But after login it hangs.
Until that’s fixed and I can’t know if it ever is, how can I just go back?
Hi, sorry things went bad for you. I was involved in the beta, even with my real Home Assistant instance and many, many integrations. The beta did not reveal big problems.
You say you recovered a backup. That is great. Do you have that backup on another system then the one you installed HA on? That is hugely important before we try to do things that might help but also have a chance of making things worse. It is your lifeline, it should never be only on the same system that you are trying to protect from breaking.
If yes, then there are things we can try. But before that we need some more info.
What version did you have before you updated to 2025.2.2? If it is 100% likely the update caused it then the safest way to roll back is the backup. There are other ways too, but they are not recommended if the rollback is to a much lower version and we’d need to know to what version you’d have to roll back to.
Did you check for breaking changes of the updates and the versions you skipped for anything relevant to you? They are in the release notes and might hold a clue to what broke down. If you used HA for many things, there are also many (custom) integrations that may cause this. For instance Alexa Media Player is notorious for causing trouble on updates.
Besides rescuing the backup, did you also manage to get hold of the logfiles from HA? These may also hold the answer to what is going on.
Yes I managed to scp it to another PI, but after rebuild how do I get it back, USB sticks aren’t supported, if I have to reconfigure all the networking before I restore I’m wasting hours…
How do you restart HA in safe mode with no interface only SSH?
What log files, what am I looking for and where are they?
Remember the system is working MQTT, sensors, lighting etc
The only failure is the GUI after logging in you’re lock in “Loading Data” that’s not systemic it’s isolated and specific probably configuration. It’s routing between virtual and physical or a setting that been changed or a black hole Web service with no response or event capture.
It isn’t Zigbee or any of the interfaces, they’re working as demonstrated by operational lighting etc. the Web GUI is foobar and that’s been done in the Core 2025.2.2 update because it was working then it wasn’t and the update is the only difference! So what did “you” (the developers) do? Something change from that update.
I’ve just been notified by Home Assistant that something has occurred based upon Zigbee monitoring and I can’t get on to handle it? I can’t turn kitchen equipment on, I can’t log health records… All because an update authorised by you created by you has locked me out of my system… and you don’t have a reasonable reversal procedure other than rebuild, which is all tomorrow gone and I still can’t update without knowing it won’t brick my system again! How is that a solution?
That is in the documentation I linked to, using ssh to access the the Home Assistant command line interface.
You were asked to do a reboot via ssh before, and the command for safe mode is almost the same:
ha core restart --safe-mode
The logfiles are in the config folder they are called home-assistant.log or home-assistant.log.1 The .1 is the one from the previous reboot. In them Home Assistant logs errors, so anything in there that may explain what is messing things up.
But the most important questions remain unanswered. What version did you update from? What is in the release notes for updates since then? If you do not know the version number, do you have some idea when the last update was installed before this?
Now I get a Safe Mode notification popup but still see “loading data” underneath it, and when you “OK” it it still sits there until the Safe Mode warning pops up again… I’m guessing through a timeout as what it expect to happen doesn’t!
Silly question maybe, but have you cleared the browser cache / tried a different browser / used private browsing to see if it is a caching or browser compatibility thing?
I’ve tried Firefox, Brave, Chrome from several locations plus a couple of others. Yes I’ve clear data and history resetting to Virgin state.
Your App which I assume embeds a browser doesn’t work either…
Before you ask it’s all Linux based and the tablets and phone are Android. one of the tablet is used only as the Hone Assistant interface and is locked into one page, that it can’t access, which is why I can’t use any kitchen equipment right now…
We fully understand the panic. And personally I did not take offence because I understood. But remember, the forum is not an official Home Assistant helpdesk. We are just users like you, doing this in our spare time. So “Your app” as you for instance said is not Our app .
As for the future, I do have a few vital suggestions that will prevent you much pain in the future:
Don’t lag behind too far with versions.
Read the release notes of every monthly update for at least the breaking changes.
Use the new backup features to automatically store backups in a remote location. Either via a Samba Share, Nabu Casa, Onedrive or Google Drive. If you don’t and your SSD is fried, you are toast.
Install the Samba Share Addon for easy remote access to the config and backups of your system
Install the file editor addon or Studio Code Server for easy editing of your config
Do not rely on HA for medical nessessities. HA is great, but no where near that great.
Plan for this event. Make sure you have ways to control things without Home Assistant.
Read the release notes of every monthly update for at least the breaking changes.
I always read them but often they’re just the same as the previous release.
Use the new backup features to automatically store backups in a remote location. Either via a Samba Share, Nabu Casa, Onedrive or Google Drive. If your SSD is fried, you are toast.
I have FTP and Samba installed and active during use only. My NAS is ready only to Home Assistant and a remote Python FTP and processing into a MySQL/Mariadb in progress.
Install the Samba Share Addon for easy remote access to the config and backups of your system
Done
Install the file editor addon or Studio Code Server for easy editing of your config
File editor already installed
Do not rely on HA for medical necessities. It is good, but no where near that good.
not necessities records. I’ve lost about 6 hours, better than I was worried about. But that’s me - I stress too much…
Plan for this event. Make sure you have ways to control things without Home Assistant.
I’m usually belt and braces but the backup systems are in development and this caught me out of the blue. I do have october-december archived, as of the 15th I’ll have January too. I will be redesigning the interfaces to accommodate buttons and switches in the future.
Once done I reduce the recorder retention back to 10 days from 90. But that’s a few months from now.
I’ve got thermal curtains to put in yet.
I don’t mind telling you this scared the hell out of me, and I will revise my dependence and procedures around a new paradigm.
Again thank you for your help and support, it is much appreciated.
that’s why I dedicate a computer to a task. A Pi4 4GB runs my HA, a media PC runs the backup MySQL and my Database/Apache server is 3x as powerful running Mariadb and MongoDB, I have Pi-Hole on another PI4 4GB plus Node-red and another backup Pi-Hole on a Pi Zero2 W.
You see what I mean about belt and braces? I’m just unusually vulnerable to HA at this time.
The database is not important enough for me to warrant the additional cost in hardware and energy use. I consider it part of the HA setup. The config was just plain wrong, I told it that it could use more memory than needed due to a typo.
.