I am running HA on a server with ECC memory. Generally speaking it’s rock solid, so think I can exclude HW errors right from the start.
It’s running in a VirtualBox image, downloaded from here: https://github.com/home-assistant/operating-system/releases/download/12.4/haos_ova-12.4.vdi.zip
It has 4Gb memory, 4 CPUs, and 20 Gb of disk space, provided by RAID over SSDs. It’s solid.
Still, looking at the console of the VM, I am seeing this:
I started looking since I was seeing errors in my ha core logs, where I looked since all my statistics have disappeared… the logs mentioned unrecoverable DB corruption.
The kernel panic seems to be happening while working with ext2 (the file system), so I also expect a corrupted file system.
Where does the instability come from? I was not doing anything extravagant.
No custom kernel drivers, no fancy hw addons. Just Home Assistant sitting there, communicating with stuff via TCP/IP.
What kind of kernel does HAOS ship with? I would expect that they have chosen a well tested, stable and reliable source, wouldn’t they?
I have spent a lot of time setting this up, and now all my previous stats are lost, maybe even my configuration?
What is going on here?
I don’t think it’s my HW, since again, this is a server running on ECC memory, with a bunch of VMs, and nothing else is complaining, just Home Assistant…
To say I’m not happy about this would be an understatement.
Can someone please offer any explanation about what is going in?
