Warm boot with Linux kernel => 3.0 locks or fials

[Perry G]

I have been seeing this for a while but didn't know what to post before. It happens when trying to do a warm boot of the guest, but not always. Perhaps 80%+ of the time.

It will either go through the shutdown part and then just freeze before Grub shows or if it does make it past the Grub window it shows softlocks on the CPU. I really don't know what to provide but I do see it showing in the guest log.

This can be followed with a shut down (cold boot) and the guest will boot properly. I think this started happening around the time the 3.* kernels started being used.

Let me know what else you need and I will provide it.

[Perry G]

A litle more information. While testing I notice that on a warm boot and hang, you can inject a reset which does reboot the guest again but it does not solve the issue. I did this several times and after the logo screen and just before the Grub menue shows the guest hangs. Finally resorting to a power down then start and the guest boots normal.

[Perry G]

Further investigation shows this is not only happening in SMP but also i686.

[CoreyW]

CentOS 6.2 x86 kernel panic screen

[CoreyW]

A very similar problem seems to happen when cold booting a CentOS 6.2 x86 client on several different Windows 7 x64 hosts (DELL, HP laptops and desktops). VM with both Guest Additions installed and not installed exhibit kernel panics on cold boots just past the grub menu.

Sometimes, if the VM is shutdown and restarted repeatedly, it will boot properly - not all the time, though.

Attachment submitted showing kernel panic.

[CoreyW]

CentOS 6.2 x86 failure log

[web_knows]

Similar problem here with a few differences.

VirtualBox version: 4.1.12.
Host version: Windows XP SP 3
Guest version: Linux CentOS 6.2 (kernel 2.6.32)

After installing CentOS, rebooted the VM and it worked fine several times. Then, decided to install the VirtualBox Guest Additions for Linux bundled with that version of VB. Since then, I can't boot the VM anymore and it is failing with a kernel panic as showed on the screenshot attached.

[web_knows]

CentOS 6.2 kernel panic after VBox Guest Additions install

[deizel]

We ran into the ​same issue - solved by switching chipset emulation from PIIX3 to ICH9.

VBoxManage modifyvm "VM name" --chipset ich9

Does this solve the issue for you guys also?

[Frank Mehnert]

Please don't do this. ICH9 is experimental and will probably show other bugs.

[Perry G]

Replying to deizel:

We ran into the ​same issue - solved by switching chipset emulation from PIIX3 to ICH9.

VBoxManage modifyvm "VM name" --chipset ich9

Does this solve the issue for you guys also?

Sorry, I have not tried the ICH9 chipset in an attempt to work around this issue and wonder why I would have to. Set aside the fact that I have had extremely bad experiences with guests using ICH9, the additional warning when enabling it suggest that the experimental nature of ICH9 is not well supported (yet).

[Frank Mehnert]

Actually I think we found and fixed this problem in current trunk. The fix is in r41256.

[Perry G]

Thanks Frank,

I'll build out today and let you know.

[Perry G]

r41256 Seems to have fixed the issue.

Reboots actually reboot. Thank you!

[Frank Mehnert]

Thank you Perry for the feedback!

[Ralph Schmieder]

CentOS 6.2 32 Bit crash

[Ralph Schmieder]

CentOS 6.2 VBox Log file

[Ralph Schmieder]

CentOS 6.2 Configuration File

[Ralph Schmieder]

I'm seeing exactly the same issue running 4.1.18 and kernel-2.6.32-220.17.1.el6, the current kernel of CentOS 6.2, 32 bit version. The OS / kernel was running w/ a kernel from CentOS 6.1. It was painfully slow compared to a 2.6.18 kernel (b/c my Atom CPU lacks virtualization features and thus the timer calibration failed) but it was at least running w/o crashing. After upgrading to latest CentOS 6.2 packages and kernel the OS crashes (see the attached files centos6-oops.png, VBox.log and centos6.vbox).