VMs freeze with new Ryzen CPU

[Swashssk]

Hi,

some months ago I built my new PC with a Ryzen 9 3900X 12 Core Processor.
Since the very first day I have massive problem with running VMs.

If I run Linux VMs, it freezes every few seconds for some seconds.
Distribution doesn't matterr. This happens on Debian, Ubuntu, Arch, Fedora, ...
This doesn't happen with my Windows 7 VM.

In a Debian VM I get these messages every few seconds.
< can't post links: You must be a member for 1 days and have 1 posts before you can post urls.>

I never had problems like this with my older Intel CPU.
So my assumption is, that it has to do with this Ryzen CPU...

- Host OS is Windows 10.
- Hyper-V is disabled.
- Power Saving Plan I choose either "AMD Ryzen Balanced" or "AMD Ryzen High Performance".

I've already tried to disable Hyper Threading in the UEFI but this didn't help.

I'd be very thankful if anyone could help me solve this problem.
I can't use VMs on that machine since weeks and it gets frustrating.
Currently I run VMs on my old machine but that's not really a comfortable solution.

[scottgus1]

Pick one guest. Start the guest from full power off, not save-state. Run until you see the problem happen, then shut down the guest from within the guest OS if possible. If not possible, close the Virtualbox window for the guest with the Power Off option set.

Please right-click the guest in the main Virtualbox window's guest list, choose Show Log. Save the far left tab's log, zip the log file, and post the zip file, using the forum's Upload Attachment tab.

[Swashssk]

Hey,

Thank you for the quick reply!

I started a VM and left it there for some hours now. The error message didn't show up.
But I didn't play around with it since I was AFK.
My PC was on 35% CPU all the time, maybe that was the point why there was no error.
I have the assumtion that it happens only/more when the PC is in idle or low CPU usage.
Maybe has to do with power states?

However, thank you very much for the tip about the log.
I'll keep an eye on it for the next days and when I get the errors next time, I'll upload the logs.

Meanwhile, in the attachment is a screenshot of the error messages (Idle Debian VM) that I wanted to link in the first post.

[scottgus1]

Even a working guest's log might help. I'm looking to see if Hyper-V is really off: there's lots of Windows services that use it, see I have a 64bit host, but can't install 64bit guests

[Swashssk]

I've started a Debian VM, logged in (tty, no gui on that machine) and then shut it down via 'poweroff' command.
While logged in, I notices that the cursor didn't blink in a regular intervall.
I'm not sure if that's important information...

Okay, in the attachment is the log.

[scottgus1]

Well, it isn't what I thought it was. Your AMD-V is working well, and no Hyper-V interference. I'll have to leave this one for the gurus to pore over...

[Swashssk]

Okay. Thank you for trying to help.
Have a good time.

[moumoku]

Okay. Thank you for trying to help.
Have a good time.

If you have an open ticket, would you mind linking me to it? I'm having the same issue, but it seems to be exclusive to Windows guests.

[multiOS]

@Swashssk

I'm sure you've already 'googled' those rcu error messages and found the many threads about incompatibilities between Ryzen CPUs, various Linux kernels and systemd, all of which go back over at least the past 2 years; and none of which seem to come to a clear conclusion/remedy.

Perhaps it's time to try some none-destructive experimentation with one of your Linux VMs. From your log, you seem to be allocating a large amount of resource to the VM, although those resources are affordable, so should, theoretically, cause no problem. Not sure if you've already tried this, but I'm curious about what happens if you limit the initial resource allocation to something more 'conventional' for a linux based virtual machine, and also closer to the generally recommended resource for an efficient physical linux setup? I would suggest starting with 2 CPUs and 2 GB RAM for a Debian System and testing if the same warning/error messages still appear, then scaling up the allocations gradually to continue checking for changes in system behaviour.

Don't have a Ryzen system myself, so don't have any first-hand knowledge to contribute.

[Swashssk]

Hey,

I just cranked up the CPU cores for taking the log since I have the feeling that if I add more cores, the errors appear more frequently. But not 100% sure about that.

Normally I have it at 2 cores and 4 GB ram.
But I still have the problems there.

Yes I already googled a lot but couldn't find a proper solution.
So posting in this forum was kind of my last hope.

[Swashssk]

I just installed an ArchLinux system in a VM and there it is extreme...
Long freezes every few seconds.

In the log I can see the following:

Code: Select all

00:27:47.831124 E1000#0: Interrupt attempts: 924
00:27:47.831155 E1000#0: Interrupts raised : 807
00:27:47.831284 E1000#0: Interrupts by RX  : 150
00:27:47.831286 E1000#0: Interrupts by TX  : 98
00:27:47.831288 E1000#0: Interrupts by ICS : 577
00:27:47.831291 E1000#0: Interrupts by RDTR: 0
00:27:47.831293 E1000#0: Interrupts by RDMT: 0
00:27:47.831295 E1000#0: Interrupts by TXQE: 0

I'm not exactly sure what this means but 'interrupt' doesn't seem to be good.

[scottgus1]

I have seen new processors induce some tweaks from the devs to the Virtualbox code, also a Linux guest OS boot fine on my Intel host but not on my AMD host requiring an in-the-Linux-guest grub tweak to get the guest to boot.

I think your issues should be posted on the Bugtracker, with suitable logs and a link to this thread, so the devs can be alerted to a possible new processor glitch.

Also, since the problem happens with Linux but not Windows, troubleshoot as if Virtualbox were not involved, and see if a grub switch or such, common to all the distros that have this problem, might be needed.