Duplicate attempts to destroy instance when calling os-stop and os-start

Description
===========

The libvirt driver's implementation of power_on currently uses _hard_reboot under the hood to recreate the underlying domain, volume connections and vifs. As part of this we attempt to destroy the domain but in the context of power_on that has already been handled by an earlier call to power_off. AFAICT the existing exception handling within destroy stops this ever being an issue but it is still inefficient and should be removed.

Steps to reproduce
==================

* Stop an instance
* Start an instance
* Note the duplicate attempts to destroy the instance

Expected result
===============

No attempt is made to destroy the instance while starting it.

Actual result
=============

An attempt to destroy the instance is made and while this doesn't result in an error it is inefficient.

Environment
===========
1. Exact version of OpenStack you are running. See the following
  list for all releases: http://docs.openstack.org/releases/

   master

2. Which hypervisor did you use?
   (For example: Libvirt + KVM, Libvirt + XEN, Hyper-V, PowerKVM, ...)
   What's the version of that?

   libvirt

2. Which storage type did you use?
   (For example: Ceph, LVM, GPFS, ...)
   What's the version of that?

   N/A

3. Which networking type did you use?
   (For example: nova-network, Neutron with OpenVSwitch, ...)

   N/A

Logs & Configs
==============