Home assistant Crashing, reboots to safe mode

Installation
1

Woke up to my HA not responding. after a hard reboot it will start in safe mode with a notification about an invalid config. If I can figure out how to attach my logs I will

1 Like
2

Did you upgrade HaaSOS to 14? I was having problems with the services not responding every 1 hour or so until I rolled back to 13.2. Even with the latest HA 2024.12 and 13.2 I have not had the problem. I think something is broken in 14 as there seems to be several of us seeing hanging service issues. Try to go back to 13.2.

3

I must have. see below

  • Core2024.12.0
  • Supervisor2024.11.4
  • Operating System14.0
  • Frontend20241127.4

Any idea how to roll back? I dont see the option to roll back os in my backups?

Thanks,

4

you cannot do it from the backup since that contains the addons and the HA core itself. From the HA commandline you can type
ha os update --version 13.2 to roll back.

5

@Mildtuna what hardware are you running HA on?

6

a beelink pc as of 2 months ago

7

Mine is doing this however I’m running mine in Docker. My container logs show:

services-up: info: copying legacy longrun home-assistant (no readiness notification)
s6-rc: info: service legacy-services successfully started
Traceback (most recent call last):
  File "<frozen runpy>", line 198, in _run_module_as_main
  File "<frozen runpy>", line 88, in _run_code
  File "/usr/src/homeassistant/homeassistant/__main__.py", line 227, in <module>
    sys.exit(main())
             ~~~~^^
  File "/usr/src/homeassistant/homeassistant/__main__.py", line 174, in main
    args = get_arguments()
  File "/usr/src/homeassistant/homeassistant/__main__.py", line 95, in get_arguments
    default=config_util.get_default_config_dir(),
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: module 'homeassistant.config' has no attribute 'get_default_config_dir'
[14:21:43] INFO: Home Assistant Core finish process exit code 1
[14:21:43] INFO: Home Assistant Core service shutdown

Not sure if this is related?

Edit: I probably should have mentioned. Downgrading my HA version gets it working again. If I go back versions 2023.6 it starts up fine.

8

Hello Jesse Soonias,

That points to a problem in the configuration.yaml or one of the other yaml files feeding it.

You have safe mode, try to do a check config and see if you can find it and fix it.

Start with anything you have edited since the last time prior that you rebooted…

9

was able to roll back to 13.2. will check back soon for stability.

Thanks!

1 Like
10

same here. Let me know if the downgrade worked for you also.

1 Like
11

FYI Alexa Media Player custom integration has failed again this release, if you have that disable it.
The Devs are considering abandonment at this point as it is borderline unrepairable.

12

No issues since downgrade. If you update and all goes well maybe come back and let me know! Thanks!

13

I don’t have this one enabled. Thanks for the heads up!

14

Also on a Beelink, and downgrading seems to have resolved it.

15

Hey @salscode , did you ever try upgrading?

16

Hi @mildtuna and @salscode,

Was curious what Beelink model are you running? I just got a brand-spanking new Beelink GTi14 Ultra AI PC Intel® Core™ UItra 9 185H (because it’s a pretty box, beast-mode specs, was on sale and I planned on trying to get Frigate to run better for me). I am coming from an Intel NUC NUC12WSKi7 and was eager to get off it since Intel killed NUC.

I’ll just say, it’s been a month of frustration. And newer shinier machine != better performance necessarily. My first attempt at installing HAOS on this monster wouldn’t boot. I found that I had to disable VT-D and X2APIC in the BIOS (in addition to Secure Boot per documentation). Unfortunately my Frigate addon install went from bad to worse. The FFMPEG CPU usage went to CPU-only and performance isn’t great. My Coral-based detections remain fine, but FFMPEG CPU usage is pretty bad.

So I figured I’ve got some time off end of this month, I should play and tinker with some things. So I got the bright idea to switch from HAOS to Supervised Install on Debian Bookworm. The warnings and the community say this is really an advanced config and they are sure right. But I’m adventurous and am learning a lot which I love (I’m don’t know Linux well and this was really fun to learn). It’s not been any better and stability is way worse. For example I’ve learned how to py-spy and am seeing trends where my templates are causing a lot of performance issues which cause HA to just churn and spin even on this brand spanking new hardware. I do see an inbound jinja update in Core 2025.1.

Turns out that the GPU (which would fix my Frigate issues) in the Beelink GTi14 is Meteor Lake H, which according to the Linux Kernel version history is not supported fully until 6.7 or 6.9 (without backports). As of right now I have no /dev/dri. The problem is Debian Bookworm (the only SUPPORTED Supervised install HA) uses the 6.1 kernel (again without backports) and HAOS is 6.6 currently with HAOS 14.1. Looks like Debian Trixie (the next version) is due out in 2025 and will likely use Kernel 6.12, which might fix all my issues, but no telling when the Home Assistant folks will have the time to test and approve Trixie for Supervised Installs.

On top of all that, stability of my HA has gone way down. It’s been kinda a mess and my partner is getting tired of Home Assistant being down all the time. My choices at this point are:

  1. roll back to the Intel NUC and downgrade to 13.x HAOS as you all have done, but I’m basically back to square one and have this shiny new BeeLink doing nothing (I guess I could put Ubuntu w Docker and dedicate it to Frigate, but that seems like a bit of a waste.)
  2. Switch the Bee-Link back to HAOS 13.x and wait it out (with disabled VT-d and X2APIC), live with Frigate performance, hope HAOS gets to a Kernel 6.12 but who know when…
  3. Wait for supervised Debian Trixie, that may be a next year, who knows?
17

You could run Proxmox on this system and never look back. Run any OS, Docker or other workload at any time.

Unfortunately Beelink is a brand which should be avoided (I own several of their systems).

Their weak points are:

  • fans, which they do not last very long
  • not enough compatibility testing prior to release
  • no long term support with FW updates or parts
1 Like
18

I have the Beelink MINI S 12 PRO, been running it for a little over a year. Upgrading from 13.x to 14.0 is the first real issue I’ve had with it, it has been a lot more reliable than the Raspberry Pi I was using before.

1 Like
19

I actually just changed hardware to a belink from an odroid to hopefully get a bit more memory and better overall performance. I am hitting a similar issue when i hit restart after installing an updates it fails to restart. The observer is saying up and running, but can’t access anything. I reboot the host and it comes back up.

Trying to make the call of going back to the odroid or install proxmox on the belink. I didn’t really want another set of services to manage and upgrade which is why I was staying away from proxmox.

1 Like
20

Yeah, I am with you. I am not opposed to doing proxmox, but I am genuinely curious at learning the absolutely most about Linux, Debian and Docker which I am enjoying a great deal. Each issue that I run into helps me learn just a little more about kernel versions, backports and all that. I don’t mind suffering a little so that I can have the most information to make the most informed decision.
If I get to a point when I want to do proxmox, I’ll be digging just as deep to learn it

One thing I found recently is that the Beelink I have uses the Intel i222-V NIC has a lot of support issues on Linux (with ASPM/EEE and firmware). I just put a USB NIC in, and so far, it’s been more stable.