ProxMox host freezes, needs hard reboot

[vale.maio2]

Hi all, I'm reinstalling proxMox on a PC, and simply put it freezes after a few hours staying idle. The machine I've installed it has been my desktop Windows PC for a couple of years, during which I've never seen anything like this happening. Here are the hardware specs:

• Motherboard Gigabyte GA-A320M-S2H AM4 socket, with latest BIOS update version F56g (December 2023)
• AMD Ryzen 5 1600X 6C/12T
• 16GB single bank DDR4 non-ECC RAM, 2666MHz
• 2x 480GB SSD in ZFS RAID1
• 3x 8TB 7200rpm HDs in RAIDZ (RAID5 equivalent in ZFS)

This ProxMox server is currently running a single Linux Mint 21 VM, with 10GB of RAM assigned and 8 CPU threads.
All the hardware is in good order, and it has been used personally for a couple of years, as I mentioned. I did run a memtest, and the RAM came out clean, no errors. Checking journalctl, there's nothing much happening righ before the crash, except some HDD smart values changing. When booting, I can see the following error:

Feb 21 07:28:12 server-proxmox-AMD kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Feb 21 07:28:12 server-proxmox-AMD kernel: ACPI: \_PR_.C000: Found 2 idle states
Feb 21 07:28:12 server-proxmox-AMD kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Feb 21 07:28:12 server-proxmox-AMD kernel: ACPI: \_PR_.C002: Found 2 idle states
Feb 21 07:28:12 server-proxmox-AMD kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Feb 21 07:28:12 server-proxmox-AMD kernel: ACPI: \_PR_.C004: Found 2 idle states
Feb 21 07:28:12 server-proxmox-AMD kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Feb 21 07:28:12 server-proxmox-AMD kernel: ACPI: \_PR_.C006: Found 2 idle states
Feb 21 07:28:12 server-proxmox-AMD kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Feb 21 07:28:12 server-proxmox-AMD kernel: ACPI: \_PR_.C008: Found 2 idle states
Feb 21 07:28:12 server-proxmox-AMD kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Feb 21 07:28:12 server-proxmox-AMD kernel: ACPI: \_PR_.C00A: Found 2 idle states
Feb 21 07:28:12 server-proxmox-AMD kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Feb 21 07:28:12 server-proxmox-AMD kernel: ACPI: \_PR_.C001: Found 2 idle states
Feb 21 07:28:12 server-proxmox-AMD kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Feb 21 07:28:12 server-proxmox-AMD kernel: ACPI: \_PR_.C003: Found 2 idle states
Feb 21 07:28:12 server-proxmox-AMD kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Feb 21 07:28:12 server-proxmox-AMD kernel: ACPI: \_PR_.C005: Found 2 idle states
Feb 21 07:28:12 server-proxmox-AMD kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Feb 21 07:28:12 server-proxmox-AMD kernel: ACPI: \_PR_.C007: Found 2 idle states
Feb 21 07:28:12 server-proxmox-AMD kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Feb 21 07:28:12 server-proxmox-AMD kernel: ACPI: \_PR_.C009: Found 2 idle states
Feb 21 07:28:12 server-proxmox-AMD kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Feb 21 07:28:12 server-proxmox-AMD kernel: ACPI: \_PR_.C00B: Found 2 idle states

I did disable C-state from the BIOS menu, and I've installed the latest AMD microcode from the non-free-firmware Debian repository. I've also added this line to the grub:

GRUB_CMDLINE_LINUX_DEFAULT="quiet tsc=unstable trace_clock=local acpi_osi=Linux pcie_aspm=off"

as I've seen people online with the same issue as mine solving it with that. Needless to say, none of this worked so far. I'm attaching a pastebin of the relevant journalctl, in the hope someone smarter than me can point me in the right direction.

 

[jsterr]

Hi can you try:

processor.max_cstate=0 intel_idle.max_cstate=0
check cat /sys/module/intel_idle/parameters/max_cstate to see the maximum allowed cstate

Greetings Jonas