This makes slapos node instance slower, especially when there are many kvm partitions on the node.

We don't really need to sleep because if qemu is not ready, promise will fail and succeed on next run.