Random freezes due to host CPU type

[Decco1337]

Hey guys,

I am facing random freezes of the whole on Proxmox 8.2.4 with the newest kernel (also older kernels), the newest microcode updates, disabled c-states, Eco mode on Ryzen 9 7950X. I found out that the issue just occurs if VMs have the host CPU type and especially with Windows VMs. I have to reboot the host manually that it cames back, I cannot type anything, it just don't respond anymore. Has anyone the same issues and a solution for it?

I also replaced the whole hardware, without any changes.

 

[Decco1337]

Updated: Seems not to be only windows related. Also got the same issue with Linux VMs (Debian, Ubuntu…)

 

[madcityservers]

Had the same issue. What I ended up doing is using the script here https://tteck.github.io/Proxmox/#proxmox-ve-processor-microcode and used the bottom option. Have not had issues for over 1 week now.

 

[Decco1337]

Had the same issue. What I ended up doing is using the script here https://tteck.github.io/Proxmox/#proxmox-ve-processor-microcode and used the bottom option. Have not had issues for over 1 week now.

Thanks I'll try that! Unfortunately a week is not long, sometimes I have no issues for one or two months :/

 

[madcityservers]

2 of my servers would freeze at least 3 times per day before I ran that script.

 

[Decco1337]

2 of my servers would freeze at least 3 times per day before I ran that script.

Unfortunately did not help... Crashes again. Moved all VMs away to my second host which has no issues and did a memtest again. No errors, it passed without erros.

The second host crashed in that time, seems still be VM related.

 

[Decco1337]

@madcityservers have you deactivated the internal GPU of the CPU or did something else in the settings?

 

[Decco1337]

Update, could saw that issue in the console after a freeze

 

[madcityservers]

Cpu issue. Had the same error on one of my Intel servers.

 

[Decco1337]

Yeah already replaced the server hardware multiple times

 

[Maximiliano]

Hello,

Please never run random scripts from the internet as root.

If you need to get the latest microcode updates for your CPU we have instructions in our documentation. Not having the microcode installed properly could explain the issue you mention.

[1] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#sysadmin_firmware_cpu

 

[Decco1337]

Hello,

Please never run random scripts from the internet as root.

If you need to get the latest microcode updates for your CPU we have instructions in our documentation. Not having the microcode installed properly could explain the issue you mention.

[1] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#sysadmin_firmware_cpu

Already did before and I checked the script, I am not dumb. Your answer is not helpful… Still having issues, seems to be Proxmox related. You should fix your software!

 

[Leask]

I experience the same issue randomly. I tried using standard Debian, and it runs fine on the same hardware.

 

[Maximiliano]

If you are experiencing "random" freezes, I would recommend installing the CPU microcode and if that does not help, then I would recommend to try different kernel versions. Starting with the latest version available in the repositories (6.8.12-4 as of today), then the opt-in 6.11, and finally 6.5 if none of the previous ones help.

Debian comes with an older kernel version which might have not had the same bug.

You can find instructions on how to pin a kernel version at our documentation [1].

[1] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#sysboot_kernel_pin

 

[fiona]

Hi,
regarding the original report with Ryzen 7950X, there was a recent fix in the kernel:
https://git.proxmox.com/?p=pve-kernel.git;a=commit;h=d5f7dca7bfe58a3e2894313ece3261656a49b1b3
included in kernels >= 6.8.12-4