device I/O errors panic with multiple CPU

[waldorfm]

Hardware: Mac Pro 2009 8-core, 10.6.8/ 64 bit, 24 GB RAM, 3 TB disk space.
VM: VBox 4.3.16, 4 CPU, 8 GB RAM, SATA and SAS devices. Oracle Linux 6.5 x64

I'm setting up a virtual machine to install Oracle Database 12c. The Database configuration assistant is repeatedly causing the following errors:

Read-error on swap-device
EXT4-fs (dm-0): previous I/O error to superblock detected
Buffer I/O error on device dm-0, logical block 0
EXT-fs error (device dm-0) __ext4_get_inode_loc
comm kworker/u:5: unable to read itable block
failed command: WRITE FPDMA QUEUED
kernel panic - not syncing

I've been playing all weekend trying several different kernels and settings. I tried 2.6.32, 2.6.39 and 3.8.13, but all produce the same errors. The only thing that seemed to improve was to increase the timeout of the devices, e.g. echo 180 > /sys/block/sda/device/timeout. However, although the errors at the console disappeared, the installation was still corrupt and accessing the command line and typing "ls", for example, would show "bus error".

What corrected the problem in the end was to reduce the VM to 1 CPU and the installation worked fine.

Is this a known issue?

Thanks.

[waldorfm]

Btw, I decided to give an old version of VMware Fusion 5 a try and the same setup and installation worked fine without any issues and probably twice as fast as in VirtualBox. So this should prove there's no problem with my hardware. Looks like VirtualBox storage I/O and CPU handling is not working properly.

[aeichner]

Please open a ticket in our public bugtracker and attach the VBox.log of the failing VM.

[waldorfm]

https://www.virtualbox.org/ticket/13476