Hi,

some users experience problems when running VirtualBox on Linux hosts
with multi core CPUs. For instance, from time to time the VM seems to
hang and after a short time it continues. Sometimes even the host seems
to hang. Moving the mouse or pinging the machine over network from the
outside makes the host and the VM continue until it stops again. In some
related reports I proposed to use the kernel parameter 'nohz=off'. This
seems to help a little bit.

Starting with VirtualBox 1.5.6 the vboxdrv host kernel understands a new
kernel parameter. Users who experience such problems should try this
paramter. Edit /etc/init.d/vboxdrv and look for the line which contains
'modprobe ...'. The location of this line depends on the package you
installed. Right behind the name of the module ($modname or something
like this), add

  force_async_tsc=1

After doing this, make sure no VM is running and do

  /etc/init.d/vboxdrv restart

Then check with 'dmesg' if the TSC mode is now 'asynchronous'. If so, then
you did it right otherwise please have a look at the script again. Btw,
the SVN version knows this parameter as well. Finally try to start a VM.

I would like to get some feedback if this parameter works for you or not.
Please make sure that we are talking about the same symptoms. The problem
is most probably that many CPUs stop the time stamp counter during some Cx
states (with x >= 2) or the counter values differ between the cores.

Kind regards,

Frank
-- 
Dr.-Ing. Frank Mehnert    Sun Microsystems    http://www.sun.com/

Attachment: signature.asc
Description: This is a digitally signed message part.

_______________________________________________
vbox-users mailing list
[email protected]
http://vbox.innotek.de/mailman/listinfo/vbox-users

Reply via email to