reboot-instances fails after nova-compute process reboot

[zhangtralon]

when the nova-compute process is rebooted, there may be some instances at the current host in the rebooting process.

we know that if instances' task_states are in the reboot_pending and reboot_pending_hard statuses,these instances will reboot after the nova-compute process is rebooted.
But, these instances fail due to the reboot operation using the rpc call, because the rpc call can fail due to the network impact.

There is no need using the rpc call to reboot the instances, because the rebooted instances are just at the current host.
This way not only reduces the possibility of reboot failure but also keep paces with the other operation after the nova-compute process reboots.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/170110

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/170110
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=432c34feb21b40fed1e03a8390f8aece1f956210
Submitter: Jenkins
Branch: master

commit 432c34feb21b40fed1e03a8390f8aece1f956210
Author: <email address hidden> <email address hidden>
Date: Thu Apr 2 20:41:21 2015 +0800

    change the reboot rpc call to local reboot

    Instances fail due to the reboot operation using the rpc call,
    because the rpc call can fail due to the network impact.
    Due to the rebooted instances are just at the current host,
    There is no need using the rpc call to reboot the instance.

    Change-Id: I3d435cf9ca5f722ed3fc4325e7331f29a73198d5
    Closes-bug:#1439493