Mitaka and Newton are both end of life, but looking at current master (rocky) this is still a problem most likely:
https://github.com/openstack/nova/blob/d4dbb42593893c1d1ed51a127b7183a314bcac2c/nova/compute/manager.py#L3102
Unless something magically started cleaning up the guest on a failed evacuate.
Can you tell where the evacuate failed? Was it during or after the driver.spawn() call here?
https://github.com/openstack/nova/blob/d4dbb42593893c1d1ed51a127b7183a314bcac2c/nova/compute/manager.py#L2906
Mitaka and Newton are both end of life, but looking at current master (rocky) this is still a problem most likely:
https:/ /github. com/openstack/ nova/blob/ d4dbb42593893c1 d1ed51a127b7183 a314bcac2c/ nova/compute/ manager. py#L3102
Unless something magically started cleaning up the guest on a failed evacuate.
Can you tell where the evacuate failed? Was it during or after the driver.spawn() call here?
https:/ /github. com/openstack/ nova/blob/ d4dbb42593893c1 d1ed51a127b7183 a314bcac2c/ nova/compute/ manager. py#L2906