Windows 10 guest crashing regularly

[TechJack3]

Hey to all,

So I've been having some issues with a windows 10 x64 vm crashing at random times. Since I can only dedicate 2 cores and 3 gigs of ram to the vm, I've disabled windows defender and uninstalled all the unnecessary metro apps in it hoping to create a quick running windows 10 vm. I am running Virtualbox 6.0.4 on the host and I have installed the guest additions 6.0.4 on the guest. I have also installed the virtualbox extensions pack 6.0.4 on the windows 10 x64 Host. The host had 4 cores total and 8 gigs of ram. I gave the guest 2 cores and 3072mb of ram to it. The guest is a dynamic vdi set for 100gb storage. I've attached the log from the last crash (up time just about 3 hours) hoping someone will see what might be causing the instability.

Thanks

[mpack]

The VMSVGA graphics chipset was designed for Linux VMs. How come people keep making a beeline for the only one of three options that will NOT work in their Windows VM, overriding the default in order to do it?

Switch back to VBoxVGA graphics and reinstall the GAs.

[TechJack3]

Thank you mpack for the reply! Just to clear up any confusion, the graphics controller (via gui) is currently set to VBoxSVGA. You are saying that I need to swap it to VBOXVGA? The defualt when I created the vm was VBoxSVGA.Was there something in the logs that indicated the VBoxSVGA was the problem?

Regards,

[BillG]

Correct. VBoxSVGA is new in this version of VirtualBox and is not recommended for Windows guests yet. VBoxVGA is the tried and tested controller.

[TechJack3]

Thanks BillG,

So Im posting last nights log in which the virtual machine locked up at 4:41 am. Maybe you all will see a common denominator between the two events. I am changing the graphics controller now to VBoxVGA as recommended by you both (mpack and BillG) since the machine had to be powered off anyways. I'll report back any irregularities or freezes. Thanks for helping me get this stable you both.

[TechJack3]

So after swapping the graphics controller to VBoxVGA and uninstalling and reinstalling guest additions the vm eventually froze. It ran for roughly 16 hours without hiccup. I just now checked on it and it froze about 10 minutes ago. Attaching the log in hopes someone can figure out whats hanging it up. I have been watching its resources and it doesn't run out of any prior to the crash as far as I can tell.

[mpack]

10:33:40.991760 RTC: Stopped the periodic timer
...
10:53:10.176815 GIM: HyperV: Guest indicates a fatal condition! P0=0x1a P1=0x9696 P2=0xffff9d0001a27bb0 P3=0x0 P4=0x0
10:53:10.176950 GIMHv: BugCheck 1a {9696, ffff9d0001a27bb0, 0, 0}
10:53:10.176951 MEMORY_MANAGEMENT
10:53:12.300301 VMMDev: vmmDevHeartbeatFlatlinedTimer: Guest seems to be unresponsive. Last heartbeat received 4 seconds ago

This is a fail inside the guest OS, not in the VM process. It could be a problem with the Guest Additions, or it could be nothing to do with VirtualBox at all.

What was the guest OS doing at the time of the freeze?

I would be tempted to try disabling the Hyper-v compatible API presented to the guest, as the "GIM" seems to be related to that.

[TechJack3]

It was just running deluge and a vpn client when I left it. I left task manager up on the vm last time it froze and I noticed that the cpu,ram,disk, and network usage were not pegged at all. What do you mean by "try disabling the Hyper-v compatible API presented to the guest?" It also crashed again not too long before I wrote this reply, I'm attaching the logs from it. What I noticed when I discovered the last crash was that it was at the windows log in screen and unresponsive. I'm desperate to keep this vm stable!
Thanks for all of the help!

[TechJack3]

Another update. Yesterday I restarted the Vm after the last posted crash and it ran from then on till just 5 minutes ago. Everything was running fine until I tried to open task manager and the Vm just blue screened and rebooted. I'm attaching this log in hoping someone will catch something. These crashes are getting really frustrating and I'm out of ideas.

Mpack,

Are you referring to the the paravirtualization setting under the acceleration tap in the virtual machine system settings? If so, which one would you have me choose? Right now its set for default, but under the details for the vm under system it lists hyper v paravirtualization.

[TechJack3]

Anyone able to help? I had the vm running last night after turning off paravirtuialization and as soon as I went to check on it it crashed opening google chrome. Would love to get this vm stable.

[andyp73]

00:00:05.416676 IOAPIC: Using implementation 2.0! Chipset type ICH9

This would seem to suggest that you have the chipset set as ICH9. This is experimental, meant for OS X guests only and isn't the default for Windows 10. Why did you change it?

42:12:53.706324 GIM: HyperV: Guest indicates a fatal condition! P0=0x4e P1=0x99 P2=0x3f805 P3=0x2 P4=0x80000300003f806
42:12:53.725188 GIMHv: BugCheck 4e {99, 3f805, 2, 80000300003f806}
42:12:53.725189 PFN_LIST_CORRUPT
42:12:56.472481 VMMDev: vmmDevHeartbeatFlatlinedTimer: Guest seems to be unresponsive. Last heartbeat received 4 seconds ago

There still appears to be an issue with HyperV that is causing the guest to BSOD. Interestingly, this one has a different bugcheck code to the previous one (this is 4e the previous one was 1a). These both appear to be related to memory issues within the host kernel. I'm not sure whether it is being caused by faulty RAM on the host or by a driver, installed on the host, doing something really stupid!

-Andy.

[TechJack3]

That's weird that its indicating I'm using ICH9. I just checked the chipset and its still set to PIIX3. I also disabled the paravirtualization yesterday and came home today to the vm at the windows logon screen again. I logged in and then shut it down to grab the log that's attached. Idk why I would be having mem issues on the host when its always been stable even with the vm crashing all the time.

[andyp73]

10:27:56.607615 VMMDev: Guest Log: VBoxGuest: BugCheck! P0=0x133 P1=0x1 P2=0x1e00 P3=0xfffff8034e34b380 P4=0x0
12:37:01.579670 VMMDev: Guest Log: VBoxGuest: BugCheck! P0=0x133 P1=0x1 P2=0x1e00 P3=0xfffff8041ebd9380 P4=0x0

Your guest seems to have thrown a couple more BSODs! These two are 0x133 a.k.a. a DPC_WATCHDOG_VIOLATION.

In your case, parameter 1 is set to 1 and so a driver within the guest has spent too long at IRQL DISPATCH_LEVEL or higher and the Windows kernel has panicked!

Can you create a clean Windows 10 guest or download one of Microsoft's IE and Edge Test Machines and see if that shows the same problem?

-Andy.

[TechJack3]

andyp73 /all,

I'm in the process of setting up a new Windows 10 build 1809 VM to try out. I'm letting it run the necessary Windows setup now (ie installing and configuring bloatware lol). I created it with 2 cpus again with 3072 MB of ram and a 50 GB dynamic vdi drive. All other settings are default virtualbox settings. Is there anything you would like me to do besides standing by for a crash?

[TechJack3]

Ive been running the new VM at the desktop and it hasn't crashed yet. I haven't even installed guest additions. I don't think it will show the same behavior since the VM basically is sitting idle in comparison to the last setup.
Any Thoughts?