seabios | seabios - This code implements an X86 legacy bios

This most probably means that the boot-loader is corrupt or missing. I suggest you to verify GRUB.

You can try the following:

1. Interact with the serial port console to further troubleshoot.

2. Attach this disk or its snapshot version to a new instance as an additional (none-boot) disk and try debug it.

3. Re-installing GRUB on the desired partition would help along with this guide about GRUB config.

   I am afraid it will not be easy to apply a mitigation here. Most probably you will have to make use of a fresh instance as an alternative and transfer the existing data from the problematic disk.

As you just removed the files instead of using apt-get or dpkg to uninstall the package, the package management is not aware of your manual removal, and so apt-get and dpkg still think the newest version is already installed, and so do nothing.

apt-get --reinstall install vagrant

should solve this.

I found what the issue was.

While using nested virtualization, the option -cpu host works. This advises qemu to use the same cpu format as the host, which in our case is also a VM, and which mostly will be using the host CPU format too..

The above setting works,

unless; you are using nested virtualization over a virtualbox, and trying to run qemu on the VM. Then, to make this work, we have to skip enable-kvm and the -cpu option altogether. It does make the qemu VM run slow, but it works.