What were you trying to do that didn't work?

Start a VM locally that stopped (destroyed) after switching to postcopy.

What is the impact of this issue to you?

Low, test case fails but the VM can be started eventually.

Please provide the package NVR for which the bug is seen:

libvirt-11.10.0-3.el9

How reproducible is this bug?:

100%

Steps to reproduce

1. Set up shared storage live migration
2. Start migration

    virsh migrate --live --p2p --verbose --domain avocado-vt-vm1 --desturi qemu+tcp://10.0.160.70/system --bandwidth 10 --postcopy-bandwidth 10 --postcopy

3. Switch to postcopy, wait shortly and destroy the VM

   virsh migrate-postcopy avocado-vt-vm1; sleep 0.5; virsh destroy avocado-vt-vm1; virsh start avocado-vt-vm1

4. Try to start the VM

   virsh start avocado-vt-vm1

Expected results

The VM can be started

Actual results

The VM can't be started immediately although virsh list confirms its shut off. But it will become 'runnable' after a while without further intervention.

error: Failed to start domain 'avocado-vt-vm1'
error: Timed out during operation: cannot acquire state change lock (held by monitor=remoteDispatchDomainMigratePerform3Params)

# virsh list --all
 Id   Name             State
---------------------------------
 -    avocado-vt-vm1   shut off

# virhs start avocado-vt-vm1
-bash: virhs: command not found
# virsh start avocado-vt-vm1
error: Failed to start domain 'avocado-vt-vm1'
error: Timed out during operation: cannot acquire state change lock (held by monitor=remoteDispatchDomainMigratePerform3Params)

# virsh start avocado-vt-vm1
Domain 'avocado-vt-vm1' started

Additional information

1. Caught by test case migration.async_ops.destroy_vm_during_finishphase.destroy_src_vm.with_postcopy.p2p on both x86_64 and s390x
2. As a result of destroying the VM, the Migration stops, this can happen in different ways:
   1. 'domain X not running' in stderror
   2. no specific error message in stderror, migrate returns with 1 and stderror is just a cut of list of "Migration x %" messages
   3. job 'migration in' failed in post-copy phase
3. The test case currently only considers the last of the above error exits
4. The wait time of 0.5 is really important; not waiting and waiting for 1 second didnt' reproduce the issue for me