I’ll try more while waiting for replies but wife is somewhat annoyed so I hope asking here as well makes it a little faster.
After we had the switch from summer to winter time (might be unrelated) HomeAssistant crashed and was not responsive in the morning.
Rebooting the PI5 took like 30 minutes but HA did come back online. It was rather slow but seemed to work after rebooting HomeAssistant itself.
An hour later it crashed again. Rebooting helped again.
As I had a full day ahead of me I tried to make a backup, which failed but did not diagnose further.
This morning, HA was down again and the supervisor won’t start.
In HA Cli I am getting the message
[Error returned from Supervisor: System is not ready with state: setup
I have tried “banner” which only tells me that the supervisor isn’t running and does not give me all the information it should.
Every command takes minutes to resolve in HA Cli.
I get error messages that e.g. “journal service” couldn’t be started.
I don’t know what to do… I have little to no experience with errors this deep in the system.
I run my system on a raspberry pi 5 and I am booting from an SSD.
Help is appreciated … not a single light is working at the moment
It may be that the disk has died - but, is the disk full, can you maybe get at it from another machine via samba? I had similar looking problems recently just after I had converted to running on a pc under proxmox, I thought the machine I had just bought had died, but it turned out to be the proxmox disk I was using had filled up.
And are you sure your dns is correctly set?
In order to ensure time is getting updated?
Do not use your gateway like 192.168.1.1 but use cloudflare or anything else like 1.1.1.1 then you can change back to whatever you prefer.
Looking at some github issues it seems a dying ssd can lead to this as well, so the previous comment can be the case…
I tried repair, but the client only tells me that Supervisor is busy at the moment. As soon as the system lets me, I’ll post the exact wording.
edit: The wording is the same as at the beginning.
system is not ready with state: setup
What could I do wrongly with my DNS? I am plugged into my Pi and read the output from there. DNS should only come into play when the router is in there as well, shouldn’t it?
I also haven’t changed anything in that regard, as far as I know, so I’d assume this one isn’t it, if I am not missing anything.
The SSD could be the reason I guess. I would have no idea how to diagnose that.
I do not think the disk is full. I randomly checked a few day ago and was on less than 50% of the 1TB that I have.
I tried plugging it into a pi4 but the pi didn’t boot from the hard drive and nothing happened. So I am not sure at this point if it is configured wrongly or the disk failed again. I will try to access it with samba.
Might, I sincerely don’t have enough information to say yes, I am just suggesting everything, as I believe you are quite on a rush on this issue, your answer are fast
You have an error in your first screenshot referencing systemd-timesyn, so I’ll take a stab and say it’s related.
Given the Pi5 has an RTC on board, maybe it’s trying to load using an incorrect time.
If you have a battery attached to your Pi, can you try pulling it out and restarting without it? You might need to be hooked up to the internet so it can get its time from there.
What do you mean with “battery”? The one on board of the pi itself? I can try to start without that but that requires soldering, I guess? I will have a look as well. Thanks for your input!!
my new hard drive arrived today and I tried the Pi with a fresh OS install. It works. That tells me that the Pi isn’t broken at least.
However, do you have any idea if it’s possible to try and repair the supervisor if it does not respond to
supervisor repair
I assume that together with a HA Core (?) update something went wrong somewhere. I do not know if it is possible to redownload or downgrade the supervisor.
If possible I’d like to try and save the data since my last backup is from June. I haven’t done too much since then so it wouldn’t be the biggest loss there is. (Two Backups since June can’t be loaded since they seem to be corrupted or just 1.1kb large…frustrating)
However, it is still 5 months of data loss so if somehow possible I’d love to avoid this.
edit: Turns out, even this update is only a partial one and does not restore all my stuff. How frustrating.
I have a crashed HomeAssistant to deal with. I’m creating a new post since what I’m needing help with has changed, in a way. I hope that’s alright.
With a fresh OS I was able to figure out that the pi still works but the system was corrupted.
If anything, I can only access the HomeAssistant Client console by connecting a keyboard and a screen to the pi.
I’d like to try the following:
Access backups if possible. I’m 100% I did them regularly but it didn’t save them on my NAS. I’m hoping they might still in the corrupted system instead.
Reinstall docker/container/supervisor without deleting any of the data, if possible.
Can anyone link me a source for how to do that? I don’t know any commands and the ones I found online cannot be used in the HA Cli but only when running a VM.
I wanted to stick with it for the longest time, as I wanted to also do other things with the RPI4 I was using for it. I had an RPI5 laying around for about a full year, not sure the best approach for moving everything over (inluding the things running outside of the HA Supervised on the same host) - especially since the RPI4 is a completely different CPU architecture than the RPI4.
I am also using the google backup add-on, and heard that such a full Supervised backup can be restored onto HAOS, and also HAOS has ALL of the same functionality (HACS, Addons, etc… PLUS also the Visual Studio Code Editor Addon works with it too without jumping through hoops!). At the time, HAOS was not supported on the RPI5 (it is now). Setting up HAOS on a VM would allow the host to be used for other things, but Proxmox did not work on the RPI5. I discovered QEMU/KVM on Raspbian would, so I stood up a generic HAOS on the RPI5 in a VM and and then restored a full backup of Supervised onto it - and it works beautifully.
Sorry to be long winded but it’s a long way in saying there is no longer any reason to stick with Supervsed in my opinion…
Not sure to understand the situation.
You got the new drive.
Installed an new instance of Homeassistant using the new hard drive and this new clean instance boots up correctly?
Where are the backups saved?
Were the backups on the same disk of the home assistant installation, the corrupted one?
If you made a clone of the old hard drive you will just be at the same point if something went wrong…
What do you have connected to rpi? Just the drive?
I figured that the specific problem had changed. When I am reading the forum I usually look at the titles and see if I can contribute. Also, if I am looking for a solution I try to find a post based on the title.
But as a mod has merged the topics anyway, I assume that wasn’t a good idea
