when doing live migration, there are a lot of duplicate logic to release instance resource between driver destory and post live migration

[Yaguang Tang]

when doing live migration, there are a lot of duplicate logic to release instance resources between driver destroy and post live migration. instance migrated to destination
don't need a destroy in source Hypervisor, expose a cleanup api seems better to be used.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/52777

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/52777
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=5ffa65ada6adb7e3322ae88cc6ddb03998be08ac
Submitter: Jenkins
Branch: master

commit 5ffa65ada6adb7e3322ae88cc6ddb03998be08ac
Author: Yaguang Tang <email address hidden>
Date: Sat Oct 19 23:12:02 2013 +0800

    Optimize libvirt live migration workflow at source

    During live migration, after instance migrated to destination, we
    call Hypervisor destroy to clean the instance, this can be omitted
    because it doesn't exist in source server already. This patch changes
    cleanup to a public method to driver, split instance destroy to two
    step:

    1. destroy instance in the Hypervisor
    2. cleanup instance resources(block device, network)

    Change-Id: I65e711c03fee7342b1e2bc4f9459baae7729f911
    Closes-bug: 1240820