I am running Home Assistant OS under Proxmox as a virtual machine, I Installed it with one of these scripts from the forum which downloads the vmdk, converts it to proxmox VM and installs and configures it.
Everything is working to my full satisfaction and has been so for almost half a year now. If I get a core update it updates and restarts the core container, if I get a supervisor update it updates and restarts the supervisor, same with my all my installed addons.
The only exception is when there is a home assistant OS upgrade.
It downloads and installs the update
it tries to reboot the VM
it freezes somewhere in the middle of the reboot
When looking into the proxmox GUI I can see it sitting there, using constant 25% CPU (probably exactly 1 core) forever.
When it is in this state I also cannot use the “shutdown” command in proxmox, the shutdown command itself would get stuck forever, the only way is to use “stop” to kill the VM immediately.
When I use “stop” from the Proxmox UI and then immediately “start” again it will boot normally, everything will work again and the OS version will be the updated one.
I have observed this behavior a few times now, I always hope the new version will have this bug fixed and will be able to reboot, but today there was an update from 8.5 to 9.0 and it still got stuck during the reboot.
I used this (or a similar) script I found here. What is different in this particular script, what is the different VM setting that is causing this? Maybe I can just change that and don’t need to reinstall?
Here is mine. The Difference I see is mine is connected on sata0, yours is connected on scsi0.
If I try to change that without reinstall (is it even possible?), will it still boot, or would I mess it up completely?
The script that you used to create that VM, was for an older version of Proxmox. The current script utilizes the current advancements in PVE7.
If you create a manual install of HAOS in Proxmox, the outcome would be the exact same as the current script. The script only automates the process.
Hi @tteck, I have the exact same problem that a HAOS update always results in a stuck VM and I have to (hard) reset the entire PVE. I tried creating a new VM using the steps above, and indeed the EFI Disk is different to the VM I currenty run. But after restoring the backup of the current VM, the EFI Disk is also restored to be the faulty one.
Sorry @tteck , I was already referring to that post. So when the new VM is made with the script (manually), I have EFI Disk set as local-lvm:vm-100-disk-0,efitype-4m,size=4M. After backup restore, it’s local-lvm:vm-100-disk-0,size=4M so seems the restore also reverts the EFI Disk to wrong parameters.
