Comment 4 for bug 1901739

(Answering my own question, one use-case for the back-n-forth LM is: You migrate a paused a instance; it succeeds. But then you realize: "oh, I actually need to do some maintenance work on this destination, so I need to migrate this VM _back_ to the source host again".)

        - - -

The said libvirt error:

    "internal error: missing block job data for disk ..."

... was introduced in this libvirt commit:

    https://libvirt.org/git/?p=libvirt.git;a=commitdiff;h=103a4245ae
    — qemu: blockjob: Pass job into qemuBlockJobUpdateDisk and rename it

I'm checking with the libvirt developers for some more input. As this is related to libvirt Block Layer interface with NBD.