Comment 10 for bug 1646779

I did a few checks as Jordan did on the signature of this bug, and I see that test_delete_saving_image (http://git.openstack.org/cgit/openstack/tempest/tree/tempest/api/compute/images/test_images.py#n43) is always running when the issue occurs.

The test deletes a snapshot while it's being created from a server, which temporarily pauses a server to then resume it and finally destroy it.

The instance log is for instance: http://logs.openstack.org/82/409682/3/check/gate-tempest-dsvm-neutron-dvr-ubuntu-xenial/0b05dcf/logs/libvirt/qemu/instance-00000025.txt.gz

I tried to reproduce this locally running something like a few times:

tox -eall -- '(test_delete_saving_image|test_resize_server_from_manual_to_auto|test_add_remove_fixed_ip|test_create_image_from_stopped_server|test_verify_multiple_nics_order|test_max_image_meta_exceed_limit)'

but I failed to reproduce until now. I did get a kernel panic in my VM after a few runs, but that's probably unrelated.