Guru meditation with r0-assertion failed

[Manu_b]

Hello all,

Since recently (update to 5.0.26 r108824 and /or Ubuntu kernel upgrade, can't figure out which one is the cause), I'm getting "Guru meditation" when running guest with more than 8GB allocated ram (tried with 16GB and 12GB).
The error in the log where the Guru medition occurs is :

Code: Select all

00:03:52.041361 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
00:03:52.041363 !!
00:03:52.041363 !!                 Guru Meditation -2701 (VERR_VMM_RING0_ASSERTION)
00:03:52.041488 !!
00:03:52.041986 !!R0-Assertion Failed!!
00:03:52.041987 Expression: RT_SUCCESS_NP(rc)
00:03:52.041988 Location  : /home/vbox/vbox-5.0.26/src/VBox/VMM/VMMAll/PGMAllPool.cpp(2574) int pgmPoolMonitorInsert(PPGMPOOL, PPGMPOOLPAGE)
00:03:52.041989 PGMHandlerPhysicalRegisterEx 00000004057d8000 failed with -1701
00:03:52.042068 !! ACTIVE TRAP=0e ERRCD=0000000000000006 CR2=0000000000e0c008 PC=00007f718b09aaf8 Type=0 cbInstr=ff (Guest!)

Host is Ubuntu 16.04LTS (kernel 4.4.0-36-generic) on a dual Xeon with 28Gb ram system, running Virtualbox 5.0.26.
Guest OS is Centos 7 ("server with gui").

I had this working previously on the same host but needed to re-install this guest. (did not use the guest in a while, so can't tell if it was still running after ubuntu kernel update and/or Vbox update).

Searching for this error I found ticket 5253 & 10745, but this relates to old versions. In 5253 the reporter had same issue (VM would crash with 16GB while ok with 8GB allocated). Issue was with heapmemory as I understood.

I'll attach a new/complete log file later if needed (I'm currently performing some software installation at the moment on the guest with 8GB).

Regards,
Manu

[mpack]

Post the entire log please, as a zipped attachment.

[Manu_b]

Post the entire log please, as a zipped attachment.

Here's the log from previous VM, I trashed this VM and re-built a new one (as I was unsure what was the issue and could not get it working).

I'll post an updated log for the current VM (changing memory to 16G or 12G) once I finished the software install currently running on it.

[Manu_b]

Ok, I've restarted the VM with 16GB ram instead of 8GB and again same "Guru meditation error". New log is attached.

[mpack]

RING0_ASSERTION usually means AFAIK that something else is using VT-x, or VirtualBox finds itself nested inside another VM. Do you have any software installed on your host such as system level debuggers, aggressive antivirus, another VM platform etc? Are you nesting VM platforms?

[Manu_b]

RING0_ASSERTION usually means AFAIK that something else is using VT-x, or VirtualBox finds itself nested inside another VM. Do you have any software installed on your host such as system level debuggers, aggressive antivirus, another VM platform etc? Are you nesting VM platforms?

Not to my knowledge... the host has a very standard Ubuntu installation with no other virtualization software package selected during install, and no additional packages installed recently. And according to the list of installed packages, nothing looks like that. (added list of packages as attachement).

And VirtualBox is not nested into any other VM.

For info, on the same host there are 2 other VM currently running and operating fine (one linux with 1CPU/2GB and one Win7/64 with 1 CPU and 4GB).
I also have another VM I've not used in a while (Win7/64 2 cpu/12GB) which does start and operate properly.

On the other hand I have another machine to run some VMs, and has the same setup (Ubuntu 16.04 LTS for the host). On this one I have a pair of Centos VM running with 12GB allocated (and few "smaller" vms running linux & windows OS). The installation (host OS, Virtualbox version and guests OS ) is exactly the same with the same updates applied. The only difference is the hardware which is much more recent (2x Xeons E5-2620 with HT enabled instead of 2x Xeons 5080 with HT disabled). On this host, the guests starts and operate properly.

Could there be anything I missed on the guest OS that is failing that could cause such issue? As a different guest with the same memory allocation does work properly, I would not believe it could the host OS or hardware related.

EDIT : I made another attempt, changing few settings (that seemed unrelated to me) : Pointing device changed from" P/S 2" mouse to "USB", and disabled UTC hardware clock.
Now the behavior is different : The guest would start with 12GB ram, but is irresponsive after boot (although it does not finish properly the boot sequence as I never have the logon screen). Does not respond to ACPI shutdown. But no "Guru meditation" and status is "running". Few minutes after "ACPI shutdown" with no effect, I had to power-off the VM.
Starting the same guest with 10GB instead of 12GB is fine (but initially I did not tst with 10G... 8 or 12GB but not 10).

[Pimmy]

I have the same issue: 16Gb / 4 Cores assigned to a Windows 10 VM on a Debian 8 server. Five minutes after the guest Windows 10 loads, VirtualBox crashes with this:

Code: Select all

Message from syslogd@kitchen at Feb 28 17:51:40 ...
 kernel:[65380.294170] 

Message from syslogd@kitchen at Feb 28 17:51:40 ...
 kernel:[65380.294170] !!Assertion Failed!!

Message from syslogd@kitchen at Feb 28 17:51:40 ...
 kernel:[65380.294170] Expression: RT_SUCCESS_NP(rc)

Message from syslogd@kitchen at Feb 28 17:51:40 ...
 kernel:[65380.294170] Location  : /home/vbox/vbox-5.1.14/src/VBox/VMM/VMMAll/PGMAllPool.cpp(2608) int pgmPoolMonitorInsert(PPGMPOOL, PPGMPOOLPAGE)

Message from syslogd@kitchen at Feb 28 17:51:40 ...
 kernel:[65380.294226] PGMHandlerPhysicalRegisterEx 000000014d6b3000 failed with -1701
Terminated

Attached are the log and the VM config. Please help.

[mpack]

There are very few tickets that mention this error. One suggestion was reducing the RAM allocation to the VM, in your case I'd cut it in half. Ditto on the graphics RAM allocation: reduce to 128MB.

There seems to be some confusion over the number of CPU cores. The reported CPU is a quad core, but the log says you have an 8 core host. Is it two Xeon CPUs?