Hi I'm having no end of problems with my HA install on RPI5 installed via the Raspberry Pi imager. I keep getting restarts, it looked at first like issues with Matter and IPv6 but I've removed the Matter integration and still getting issues now seeing things in the logs about containers disconnecting and reconnecting and some DNS errors and issues with pulse audio so I'm not really sure where to start to diagnose this. Any help, advise or known issues would be appreciated.
#New install with Raspberry Pi 5 frequent restarts
unborn bear
wind python
Check the supervisor and core logs.
dreamy skyBOT
wind python
You can also use the SSH addon and ha core logs and ha supervisor logs.
unborn bear
Here are the two log files, I took 500 lines of each as I wasn't sure what was needed.
This is just after the last crash, only this time it didn't auto restart so I had to hard power cut it.
It's a new RPI5 and I've tried two different SD cards. I was running HA fine on a RPI3 (I say fine but mean ok)
wind python
Have you tried in safe mode and checked with glances how the memory usage looks like?
unborn bear
Not tried that. How do I use glances? Apologies I'm still very green when it comes to Linux based systems.
wind python
It's an addon. It allows you to monitor system status, memory usage and so on.
unborn bear
I'll grab that. Just looking in Config / Hardware memory is 0.8GB / 8GB
Should I run glances in protection mode?
Just launched it and answered my own question.
wind python
I think you need to disable it for it to work. Its logs should complain if it's the case.
If safe mode helps also see here: https://community.home-assistant.io/t/2024-5-tracking-down-instability-issues-caused-by-integrations/724441
unborn bear
I'll try rebooting into safe mode now.
I'm in safe mode and have glances up. What should I be looking for? Thank you for your help so far.
Right, I've just had it crash whilst in safe mode, and with glances open, so I've managed to screen shot what was on glances before reconnection was made
Just had it crash again, what I did just spot just as it crashed was pwmfan wasn't showing then it did and crashed
I also realise that because it crashed it took me back out of safe mode
wind python
Press z to show processes but I just wanted to see if memory is getting scarce. Doesn't look like it.
wind python
I don't see a lot in the core log beside that tuya issue. Not sure if that still happens in safe mode. Can you check the core logs from the safe mode run?
unborn bear
Give me a moment, it's just crashed managed 18 minutes of uptime then.
This should cover the safe mode, however I don't appear to get it to stay in safe mode for long
wind python
Looks like it's in a restart loop but I can't see why. Let's try the good ol' hammer
ha core rebuild
ha supervisor repair
ha host reboot
If that still doesn't work we can try looking at the host logs too or purge the whole thing and restore a backup.
wind python
Shouldn't, no.
unborn bear
I'll try that probably tomorrow now as its getting late. Thank you for your help.