Ok looks we have the same problem. You sure downgrading the core will not cause more damage? Have you found an official/supported method how to do it to be sure we do what the developers count on?
I recently used a snapshot to move from an SD card to an SSD drive, as well as from a 32 bit install to a 64 bit install, and HACS was successfully restored along with the lone repository I’m using.
When using an SD card, I would occasionally (~monthly) create a clone of the card, which in addition to more current snapshots would make a restore from a card failure go quickly. Now that I’m using an SSD drive I will need to create a .img file using Win32DiskImager and save that to a different device since I don’t have a second SSD drive. That’s no more difficult than cloning an SD card. The only issue with either option is the downtime while you have the disk pulled in order to create the backup copy.
The next time it happens, and after you bring it back up, check the logs at:
Configuration - Logs
Also the logs at:
Supervisor - System - Supervisor
Supervisor - System - Core
Supervisor - System - Host
I would also check the Logbook to see if it recorded anything unusual near the time it ‘crashed’.
I wonder if as the OS grows in size it may also be consuming more power, and the issue is your power supply? I believe the Rpi4 needs a minimum of 3 amps. I haven’t checked mine since it’s running just fine, but I do have one of these I would use to see what it’s drawing:
https://tinyurl.com/ycemztur
Downgraded it to 5.8 using:
ha os update --version 5.8
I am curious to see if this will solve the problem, the downgrade itself worked just fine and HA is up and running as normal. Let’s see if it will hang again.
I don´t think so (cause I´m still on 5.8 having similar issues HASS OS becomes unresponsive / then almost unusable / and is finally dead - starting every ~ 10 hours after last HA start) but fingers crossed for you 
Can you please tell us more in detail how did you downgrade? Thanks
Make sure you have installed the addon called “Terminal & SSH”, then use either SSH from your computer or go to the addon and open the Web UI and then from the command prompt run the command:
ha os update --version 5.8
That will initiate the downgrade and restart your RPi.
BTW, I am running on a RPi 4 8Gb with a new SD Card if that is any information that might be needed.
1 Like
I have the same problem.
First I thought that the deconz applications was the problem.
At some point I couldn’t check my configfile because it was not accessible
But when I look at the log files it seems the problem is database access.
So I’m trying to setup an new SD-card and see how it goes
Good morning folks, my RPI was freezed fourth consecutive time in a row. So I am going to try the downgrading @jockesoft recommended and will tell you tomorrow if there is a change. Logs are free of any suspicious.
@Klabatzie I am not sure we have the same problem as I don’t have access to RPI at all, even ping does not work.
This might be the issue we are experiencing: https://github.com/home-assistant/operating-system/issues/1129
Stefan Agner: This issue is related to freezes with zram decompression failure, so make sure to check if you have this errors in the Host log to make sure its this actual issue.
A fix for zram decompression failure will be in 5.10 onwards.
As I never upgraded to 5.9 instead seeing those issues still on HASS OS 5.8, I think it might be https://github.com/home-assistant/operating-system/issues/1119 or https://github.com/home-assistant/operating-system/issues/1096.
And there´s no solution yet so I guess this topic is not “solved” too @ Vahaldor.
Fifth day and I am running after degrading to 4.8.
I will keep it as “unresolved” as some others experience freezing even on 4.8
I had the same problem.
After a few hours HA would become inresponsive an not accessible from all divices.
Error i got on mobile devices was “502 bad gateway”
After checking the logbook when i restarted HA icouls see that it couldn’t open the database.
“Error executing query: (sqlite3.OperationalError) unable to open database file (Background on this error at: http://sqlalche.me/e/13/e3q8)”
I’ve now made an new SD-card(high performance) with a snapshot i was able to get out of HA and everything works again as a charm!
Closing on 2 full days of uptime now after the downgrade from 5.9 -> 5.8. No problems seen at all so atleast for me it seem like there are some issue in 5.9 causing my RPi to completely freeze after ~10hours. Will stay at 5.8 since it seem more stable.
That’s the zram decompression. However, things can freeze for more reasons understandably .)
Let’s check together whether our rpis stay up after updating to 5.10
Has 5.10 been offered to you already? It´s been released (stable) but not showing up in the UI yet.
No, I dont see it as well. However, I will wait couple of days to be sure there is no other reason for my freezing than zram decompression bug in 5.9
Sixth day since the problem has revealed and second day running without interruption after downgrading back to 5.8