Host system resets (crashes) when using VMWare or VirtualBox and 64-bit guest systems

Question

I have been trying to install virtual systems on VMWare for a while now and encountered strange behaviour from my PC.

The behaviour is as follows:

• On "automatic" virtualization mode it either outputs a cryptic error message (can MAYBE give later, if I can reproduce) right on startup (before even the BIOS) or it resets the complete HOST system (blackscreen, bios...)

• If I install a Windows XP on it it works well on "binary translation" mode.

• If I try installing Linux on it, in "binary translation" mode it crashes 1 or 2 seconds after I hit enter on the GRUB selection screen (after the first page of kernel messages rolled in)

• Using VirtualBox it crashes right in the BIOS. It gave me a Bluescreen though! 0x00000101: CLOCK_WATCHDOG_TIMEOUT: a clock interrupt was not received on a secondary processor within the allocated time interval

• NEWS: I tried VirtualBox again and it did not completely crash the computer this time. It gave me a critical error and a log file: http://pastebin.com/yKZSDs91

• In conclusion, it will crash instantly if VT-x is activated. If not, it seemingly only crashed if I try to install something with 64 bits.

• Another update: Yes, it ONLY crashes when the guest is 64 bit!

• On Linux using VirtualBox, it crashes with a Kernel Panic.

What I tried:

• Reinstalling Windows (my Windows installation was quite broken so it seemed natural. Didn't work though.)
• New BIOS

What I am certain of:

• Virtualization extensions are activated in the BIOS

What my computer specs are:

• ASUS P8P67 LE mainboard, newest BIOS/EFI firmware
• Intel Core i5 2500k
• Ati Radeon HD 5770
• 16 GB Corsair 1333mhz DDR3 RAM, 4 X 4 GB

Solution 1

If you have a spare hard disk laying around you could try install a copy of Linux like Ubuntu. After Linux is installed try installing VMware or Virtualbox and see if the error presists. What this would rule out is hardware malfunction.

Another thing you could try but would be much more complex would be to run a copy of VMware ESXi and make the machine into a temporary server and once again see if the problems still presist.

Something that has just crossed my mind, what is the specs of one of you virtual machines?

Are you allocating all system resources to the VMs or are you holding back?

For example I only use 1 vCPU in my VM(s) so there is still one available for the main system, granted I am only running 1 VM at a time but you get the idea.

Solution 2

Ensure all your host's drivers up-to-date with the latest version available from the manufacturer(s) (Chipset, RAID, Video, etc.). I'd recommend getting the latest drivers directly from Intel, instead of trusting Asus to have the latest posted.

Clock_Watchdog_Timeouts are often/usually caused by a kernel-level thread (say, in a HDD controller driver) that's gone into a Spinlock. That error in combination with the fact the host is resetting tells me there's either a driver problem for a piece of the host's hardware, or a piece of hardware on the host is malfunctioning.

Comments

[Admin]

Are you using dual channel or a single stick?

[Admin]

4 sticks a 4gb in total

[Admin]

You can try VirtualBox to see if it gives the same error.

[Admin]

@JoeInternet Seems that I have to.

[Admin]

VirtualBox crashes in similar ways. Right in the BIOS for example. BUT I have finally gotten a bluescreen: 0x00000101: CLOCK_WATCHDOG_TIMEOUT. Hell

[Admin]

I got a VirtualBox log! YES!

[Admin]

@sinni800: You're using a K edition processor. Is it overclocked? VT-x may reduce the effective speed you can run the CPU at if the total number of gates per cycle that need to switch are greater (which is probably the case, particularly when nested page tables get involved)

[Admin]

@BillyONeal No, I never overclock. I bought the K because it was the same price and I MIGHT want to overclock later, but practically, I never do. But what can I do about it?

[Admin]

@sinni800: I would consider returning the CPU and motherboard for replacement if you can. If you can't... if you've already tried reinstalling the host OS I'm out of ideas. Sounds like hardware to me though. You might want to debug the crash dump with Windbg to see if it'll finger a specific driver for you: forums.majorgeeks.com/showthread.php?t=35246

[Admin]

I couldn't even get the debugging tools to install. I'll try more. Until then, replacing the hardware is the answer..

[Admin]

BillyONeal was pretty much first in suggesting to replace the motherboard, but he didn't post an answer. So I will have to give the reputation, which will end in one hour, to @Solignis, he was second fastest in suggesting to replace.

[sinni800]

Sorry, I forgot to state that a) I have had no other bad experiences with it and b) I have run a nightlong memtest before, not long ago. I think I can safely say that the ram is intact

[HaydnWVN]

As a simple 'test' just run 2 of your 4 sticks and see if it behaves better - it might be some weird problem with your motherboard, 16gb of ram and Corsair chips. Can't think of much else to try except lowering memory timings if they're specified or specifying them if they're on auto...

[sinni800]

I received the other two sticks a week ago, the problem was there before :). The timings... I never cared about setting timings. I always left them at auto.. So I don't know what I would set them to

[HaydnWVN]

Ahhh ok, tried it without your old sticks in?

[sinni800]

It doesn't really change anything. It seems the error is somewhat related to support.microsoft.com/kb/975530 since I also use virtualization when this occurs... Huhl

[HaydnWVN]

I'm all out of ideas, sorry! Certainly doesn't seem to be RAM related... Next thing i would do is try a different motherboard but that's just me - i'd still point my finger at some weird hardware compatability :)

[sinni800]

I'm sorry, no. It didn't change anything. It just added support for a few things like PXE roms, USB 2.0, ... and didn't change anything at the core of VirtualBox.

[sinni800]

1 CPU, 1 gbyte of ram. I have 16 GB available, so... Should not bother my system at all. I will try running a Linux on a seperate hard drive though.

[sinni800]

But no computer test I have ever tried has ever returned a fault. I'll be on the Intel site and download everything new now. Edit: Video drivers from ATI are the latest, Chipset (Intel INF installer) didnt install anything (was current), Rapid Storage Drivers were new ones from three days ago! Fingers crossing.

[sinni800]

After trying with VT-X enabled it crashed... In the bios, with a 0x00000124 WHEA_UNCORRECTABLE_ERROR error.

[sinni800]

Im on my fresh linux system now. CentOS Linux release 6.0 (Final)

[Ƭᴇcʜιᴇ007]

You have a hardware error: "The WHEA_UNCORRECTABLE_ERROR bug check has a value of 0x00000124. This bug check indicates that a fatal hardware error has occurred. This bug check uses the error data that is provided by the Windows Hardware Error Architecture (WHEA)." Short of spending a lot of time debugging with WinDBG and such just to find a potential culprit, your best bet is going to be to start trying replacement parts (Video, RAM, Motherboard, etc.). If you don't have the test parts, then you may need to take it to a shop.

[ianc1215]

Whats the word? Does it work?

[sinni800]

Okay I got to try it now. I started virtual box, started a virtual machine, and, KERNEL PANIC. If anyone knows how to debug anything out of the kernel crash file in /var/crash, I'd be pleased. I used the crash binary to open the crash file, resulting in a GDB instance. But now I dont know what to do. Still, seems to point towards hardware failure.. Damn. Or BIOS... Any of that

[sinni800]

I hoped it would never come that far, but it really seems that my hardware is malfunctioning

[ianc1215]

It seems like you may need submit an RMA to ASUS for the motherboard. I guess that is where the problem is. On a side note by any chance is the processor overclocked? That might lead to the problems in some way.

[sinni800]

Execute bit? I am onit.

[sinni800]

Not overclocked...

[sinni800]

Damn, no. It still crashed. This is it, ASUS. Here comes your return... :/

[ianc1215]

Ok so that is ruled out. I would possibly see if ASUS has a recall on the board or something, sounds like a hardware problem, like I said I would say the motherboard unless you have reason to believe it is another part.

[harrymc]

If you can return it, I suggest buying another model (although by the specs this is a very nice computer).

[sinni800]

Well here in Germany we have, by law, a 2 year "guarantee" that things must work. In those two years it's in the responsibility of the supplier to prove that I broke it. So if they can't prove that I broke it, they have to replace it. This means I'll only get another LE board. I'll check if I can find somebody with a similar setup, so we can switch processors. Then I will see if it's the processor or the board.

[sinni800]

Here is your bounty. I won't accept until the board is replaced and it works, though :). The bounty would have ended in one hour so I kind of needed to give it to someone. Reason why you is in question comment :)

[adeel41]

Having same issue. It only works when you assign only 1 cpu core