In previous posts, I have talked about getting VMware Workstation back on its feet again after a kernel upgrade. It also seems that VirtualBox is prone to the same sort of affliction. However, while VMware Workstation fails to start at all, VirtualBox at least starts itself even if it cannot get a virtual machine going and generates errors instead.
My usual course of action is to fire up Synaptic and install the drivers for the relevant kernel. Looking for virtualbox-ose-modules-[kernel version and type] and installing that usually resolves the problem. For example, at the time of writing, the latest file available for my system would be virtualbox-ose-modules-2.6.24-19-generic. If you are a command line fan, the command for this would be:
sudo apt-get install virtualbox-ose-modules-2.6.24-19-generic
The next thing to do would be to issue the command to start the vboxdrv service and you’d be all set:
sudo /etc/init.d/vboxdrv start
There is one point of weakness (an Achilles heal, if you like) with all of this: the relevant modules need to be available in the first place and I hit a glitch after updating the kernel to 2.6.24-20 when they weren’t; I do wonder why Canonical fail to keep both in step with one another and why the new kernel modules don’t come through the updates automatically either. However, there is a way around this too. That means installing virtualbox-ose-source via either Synaptic or the command line:
sudo apt-get install virtualbox-ose-source
The subsequent steps involve issuing more commands to perform a reinstallation from the source code:
sudo m-a prepare
sudo m-a auto-install virtualbox-ose
Once these are complete, the next is to start the vbox drv as described earlier and to add yourself to vboxusers group if you’re still having trouble:
sudo adduser [your username] vboxusers
The source code installation option certainly got me up and running again and I’ll be keeping it on hand for use should the situation raise its head again.