Home Assistant Android App crashing my HA Virtual Machine on ProxMox

Hello

I migrated recently my HA environment from a NUC8i5BEK2 (running ProxMox 7.3) to a a new hardware environment (Hardware: HunSn RJ02 with 4 Ethernet interfaces with Samsung 500 GB Nvme and Crucial 32GB RAM) running ProxMox 7.4…
The new hardware is running a PfSense VM (Firewall), my Home Assistant environment in a VM and 3 containers running: WireGuard for VPN purpose, Icinga2 for network monitoring purpose and Pi-Hole as DNS server…

Since the migration of the HA VM from the NUC to the HunSn hardware, I am experiencing very slow response time on the Android App (between 20 and 50 sec for updating a card showing the time on a dashboard) as well as a daily crash of my environment (I have to stop/start the VM machine on ProxMox if the environment is not booting by itself). I am sure this is the Android App causing the issue because I removed it from my Android devices and since than I do not have any crash anymore… I am now using a browser to access my dashboard and it is working just great : fast and stable !

I looked at journalctl from the HA VM… No error message before the boot as far as I know…

Anyone experiencing any similar issues on HA Android App ? Any solutions ? How to debug ?

Thanks

To ensure compatibility with a newer Intel Celeron, you will need to install the Proxmox 6.2 kernel and the Intel microcode.

Thank you for giving some directions

Where do I need to install this ? at the PVE level or in the virtual machine running HA ?
I presume when you say Proxmox 6.2 kernel, it corresponds to the linux 6.2 kernel… because I am running Proxmox 7.4-3 which is the latest release…

The virtual machine running Home Assistant has “Linux 6.x - 2.6 kernel” as OS type…

For detailed instructions, please refer to this guide, https://www.derekseaman.com/2023/04/installing-home-assistant-on-proxmox-quick-start-guide.html

1 Like

Updated to kernel 6.2 and latest CPU microcode… Thanks !

Maybe one additional question: It seems that when there are updates to the kernel, at the end, the process rebuilds the bootfiles/environment but it starts with kernel 15 and then discover that we installed kernel 16 and rebuilds it too and then it boots from kernel 16… Is there any way to avoid to recreate the bootfiles for the kernel 15 when we upgraded to kernel 16 ?

For the time being, I have the following:

$ proxmox-boot-tool kernel list
Manually selected kernels:
None.

Automatically selected kernels:
5.15.107-2-pve
6.2.11-1-pve
6.2.11-2-pve

Should I “pin” the kernel 6.2.11-2-pve ?

Thanks

No, Proxmox still needs the 5.15 kernel for now.
You can run apt autoremove to remove any unused kernels

If I am running “apt autoremove”, it proposes to remove “pve-kernel-5.15.107-1-pve”…

So I stopped the autoremove process based on your last comment “Proxmox still needs the 5.15 kernel for now.”

You can safely remove 5.15.107-1-pve because it has been replaced by 5.15.107-2-pve which is currently installed.

Edit: Running apt autoremove will not remove any necessary kernel packages that are currently in use by the system.