VirtualBox boot problem

[Ralf Mardorf]

On Wed, 2023-03-22 at 22:11 -0400, Little Girl wrote:
> It seems that VirtualBox is kind of like a browser in the sense that
> you can't use dusty old copies of it from GNU/Linux repositories, but
> have to keep up with the cutting-edge by getting it straight from the
> source. Otherwise, you'll get strange and sometimes impossible
> behaviors from it.

Hi,

sometimes I migrate from the Oracle branded non-OSE to the OSE or vice
versa. VirtualBox isn't reliable. At the moment I downgraded to OSE
6.1.40, because 7+ doesn't work with Windows 11 guests. Since I'm doing
this on a rolling release model distro, I can't just downgrade the
package. If VirtualBox links against an updated dependency I need to
rebuild it in a clean chroot. Building the package outside of a clean
chroot fails. VirtualBox is a PITA, but unfortunately it's de luxe
related to the Linux host and Windows guest file sharing support. To my
knowledge no other hypervisor provides this comfort. This is why I
migrate between non-OSE and OSE again and again or maintain 6.1.40 at
the moment and I'm willing to accept that VirtualBox does perform very
bad compared to QEMU/KVM and that the vbox modules can only be build for
old longterm rt-patched kernels. So I either need to reboot between the
latest rt-patched kernel and a kernel without the rt-patch or I need to
build an older rt-patched kernel. However, virtualbox cutting-edge isn't
always the best solution.

Regards,
Ralf