Nova migrate fails if you do not do stop and start instance before migration

Issue

• Configure environment to allow nova migrations to be performed
• Instances enter error state after being resumed from suspend
• The migration/resume fails with errors in the nova compute log:

2016-01-28 11:19:13.039 2402 TRACE nova.compute.manager [instance: <uuid>] ProcessExecutionError: Unexpected error
while running command.
2016-01-28 11:19:13.039 2402 TRACE nova.compute.manager [instance: <uuid>] Command: qemu-img convert -f qcow2 -O qcow2
/var/lib/nova/instances/<uuid>_resize/disk.swap
/var/lib/nova/instances/<uuid>_resize/disk.swap_rbase
2016-01-28 11:19:13.039 2402 TRACE nova.compute.manager [instance: <uuid>] Exit code: 1
2016-01-28 11:19:13.039 2402 TRACE nova.compute.manager [instance: <uuid>] Stdout: u''
2016-01-28 11:19:13.039 2402 TRACE nova.compute.manager [instance: <uuid>] Stderr: u"qemu-img: Could not open
'/var/lib/nova/instances/<uuid>_resize/disk.swap': Could not open backing file: Could not open
'/var/lib/nova/instances/_base/swap_1024': No such file or directory\n"

• Issue appears after the instance has been running for about 24 hours
• Workaround is to power cycle the instance before migrating or suspending