I recall you need to downgrade to 5.13 before you can downgrade to older versions.
1 Like
I’ll throw my hat in here too.
RPi 4 64-bit, running 6.2.
Stability has gotten so bad I have put a Kasa smart switch in front of my RPi so I can just power cycle it. Comes right back up and it seems to be dying every 1-2 days.
Having same issue
I ended with ESP8266 pinging PI and switch off relay to restart it 
one-two-three days lifetime and Pi hangs, cpu temperature about 70 degree after restart (59 before hang). If connected by ehernet Pi also kill all network activity on lan, restart helps.
I am a new user and my system is running HA OS (latest) on a RPI 4GB with a 512GB SSD, no SD Card at all. Every day or two the system starts slowing down until it becomes unresponsive. CPU and memory are all low and my ability to troubleshoot is limited so I don’t know what is causing it. All I can say is that I have spent a lot of time working in Node Red and often end up with many tabs open pointing to HASS. The only thing that seems to bring things back is to reboot the host.
Any suggestions on what to look for? And how to do so?
Just incase it helps I had the same issues as @Volume9, I have the red/white standard pi4 box which was getting really hot and killing off my network over LAN when it crashes, taking the ethernet plug out the back would instantly resolve the network activity.
Not had my pi crash for a week or so not sure if just lucky but tried a few things, which might be worth a look to see if this helps anyone.
-
Moved from Ethernet to wifi on the pi4, when I did this it no longer killed the LAN/router/network activity. But HA would still crash but at least it was just homeassistant offline (then similair process to restart the power via smart plug offline).
-
Open the pi4 case lid to allow more heat to escape.
-
Disabled ipv6 completely and just leave it to connect over ip4 via the homeassistant backend settings (same place to enable wifi).
This is all running the lastest version of homeassistant supervisor (6.2?)
I tried a lot to get it stable. And it look like I found the main cause of the HA hangs. After I disabled under Integrations The Google Cast. (found it because Google Cast was giving errors in the Log). And after that It is already almost one week stable. (Did not yet removed the automatic restart at night).
I’ll give this a try out of curiosity. Keep you posted of my experiences.
@Dotly did manage to solve your issue?
Have the exact same problem, everything works fine when we are there (summer house), making me think all is OK.
After leaving it takes a day or two before it crashes.
Happened 4 times now so I’m convinced it’s not just a coincidence.
Makes me think it could be related to some housekeeping on HA starting after some inactivity?
Brand new RPi 4 4GB w/ Samsung Evo Plus 32GB Micro SD Card.
Started with HassOS 6.6 64-bit clean install. With stock integrations/HACS and its stock integration/ frontend stuff.
Unreachable (not on the local network anymore)…hang or crash? every 12 hours or less…
Installed HassOS 4.2 (Stable) 64-bit clean install, with the same add-ons as previous… with the latest core (2021.11.4) & unsupported supervisor (2021.10.8), from:
https://github.com/home-assistant/operating-system/releases/tag/4.20
Perfect uptime. No crashes at all. Absolutely stable. HassOS 4.2 is a keeper.
Update Jan 26, 2022:
It has been 3 months? Still perfect, no crashes at all. HassOS 4.2 is definitely a keeper. Running latest core-2021.12.10 & supervisor-2021.12.2.
Had the same Problems with a RPi 4 , 1gb.
Having changed the swap size, i never had these problems again. Sometimes its slow, but normaly everyting works ok.
Perhaps this helps even with Rpi`s with more RAM ?
Here my post (and the thread)
Facing this issue once a week with a Lenovo ThinkCentre M700. It reports “rcu_sched self-detected stall on cpu” at console and the cpu usage increases until I power off/on the HA VM on proxmox. It’s really annoying…
What version of Proxmox?
Thought maybe you was using a old kernel. But you’re not. The only thing I’d suggest is to create another Home Assistant OS VM, restore HA backup, shutdown current HA VM to see if maybe the VM was misconfigured
I’m pleased to be able to finally report (almost exactly a year on) that the (RPi4 firmware) issue has been fixed in OS 7.1. Relevant entry from the release notes:
I just came across this thread after having opened one every few weeks for the last couple of years posting my logs but receiving no useful help
Have everyone’s issues been fixed? I’m on Operating System Version 7.0.rc1