crash of the entire system without BSOD VBox6.1.16

[vigilian]

Hi,

I have random crashes, especially when the idle time kicks in because of the use of LHC@Home through boinc that uses virtualbox and so a lot more of VM are loaded at that time.
It's a complete crash so what i mean is that it's not BSOD for the only one I can notice (it's a remote host on my LAN but so I'm not next to it each time).
So first question: What can we retrieve as information from that? It goes to black screen. The disk, info from the HDlED, seems to work a little bit, then the host restart by itself.

The only VMs that got to work for the moment are some whonixVM and LHC@home. So not much .
It's a relatively new installation of windows but I did change its hardware configuration since it was in in-between period.
But yes I'm aware it could be a driver problem but I really hope not and I would like to work on this assumption first because if the system works correctly for everything else, I wouldn't see why it would be that problem. Plus normally if windows is working correctly, only the drivers for actual equipment are activated.

I'm more thinking about a platform specific problem.
The actual configuration is a b550 chipset + ryzen 7 3700X +32G at 3200mhz. windows 1909 pro

I've already uninstalled vbox / reboot/ reinstalled as advised in many other cases.
There has been one change, a error message about memory access is showing now and I can't play any VMs without having this error message. So i hope it was already this problem before.
available there https://share.riseup.net/#PxiN93tK4an6L004tijKZg
I really don't think it's memory corruption because I can maxout my memory with other heavy lifting and it doesn't budge.
It could be it, I still didn't do a memory check.

It could also be a security program problem.
But I first wanted your input.
Are there any other logs I should add?

[vigilian]

No apparently there was 2 separate problem.
One of my graphic card, probably an incompatibility of chipset, who made the host reboot
and the other one is an error message among VM from the LHC@Project.
Could someone tell me what does it mean this:

Code: Select all

VBoxManage.exe: error: The VM session was aborted
VBoxManage.exe: error: Details: code E_FAIL (0x80004005), component SessionMachine, interface ISession

What's the sessionMachine component and the interface ISEssion and what could interfere with it to not be able to continue a VM because of that?

[scottgus1]

'Aborted' means the VM just stopped, without Virtualbox being able to do its normal post-guest cleanup. Error E_FAIL (0x80004005) is a general error meaning 'something went wrong', and I think SessionMachine and ISession are Virtualbox source code designations for the parts that noticed the error. Only the devs and one or two forum supergurus could say more. Suffice it to mean the VM crashed hard.

If the physical PC reboots while running the VMs, but not at other times, either the VMs are touching bad memory that does not normally get touched, or the CPU overheats, or there's an incompatibility between Virtualbox and some other software.

An overheat would not cause a BSOD, but other things might, check the Event Viewer. An intensive full memory scan and CPU stress tests can check memory & overheating.

If a BSOD happens, you'd need to pin it down using Microsoft's methods.

A guest log may help. If your host PC is not overheating, start the guest from full normal shutdown, not save-state. Run for a little while 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.

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

a lot more of VM are loaded at that time.

How many is 'a lot'?

[vigilian]

Thanks for all these specs @scottgus1

So I've changed the GPU. IT might have been 2 concurrent problems.

There was no BSOD. just flat reset. nothing in the event viewer that could suggest anything.
Only the last few ones because I've switched for a few days to 1030 nvidia just to test if it was related and apparently some of the reboot was related. And when I got back to 970 gtx it crashed several times the nvidia drivers with errors in the event viewer.
So I've bought a 2070 super and for now no reboot. So There might have been a incompatibility between the GPU and the motherboard.

Another problem that I've just discover now:
I'm using docker but also boinc with LHC@Home which is a scientific distributed project running through virtualbox.
I didn't know that WSL was running on top of hyperV API through virtual machine platform feature in windows.
For the last 2 weeks now, the project isn't detecting the SVM extension from AMD anymore because I guess of the upgrade of windows to 2004 and so it doesn't run any VM anymore.
That's what I said about lot of VM because of this project which was running several concurrent VM.

But I'm going to make another topic about it to know some wierd things about it in the logs of my VMs.

So in conclusion: it might have been an incompatibility between the GPU and the motherboard or an incompatibility between virtualbox and hyperV API for WSL2.
No overheating, all tests came back negative about overheat.

[scottgus1]

Hyper-V is known to interfere with Virtualbox. The devs are working on "peace talks" between the two sides, but it's still a work in progress.

WSL2 enabled Hyper-V, the original WSL does not. So WSL might be usable with Virtualbox, but WSL2 is not.

See I have a 64bit host, but can't install 64bit guests posts 2 & 3, and HMR3Init: Attempting fall back to NEM (Hyper-V is active)

[vigilian]

I've made the follow up discussion there: viewtopic.php?f=6&t=100789