Comment 8 for bug 1950310

summary from today's troubleshooting (for the longer form read the IRC log[1]):
* the kernel panic happens during or right after the live migration finishes so when the volume detach happens the guest is already hanging
* the oldest log we have is from 29th of Oct and it is green, and it has the same qemu and libvirt versions (qemu-system-x86 amd64 1:4.2-3ubuntu6.18 and libvirt0 amd64 6.0.0-0ubuntu8.14) as the recent runs producing the panic
* the october run uses cirros 0.5.1 as well as recent runs. The master uses 0.5.2 but we was able to reproduce the bug with 0.5.2 on stable/victoria so it is not the guest OS

[1]https://meetings.opendev.org/irclogs/%23openstack-nova/%23openstack-nova.2021-11-12.log.html#t2021-11-12T13:17:46