redundant image GET calls during rebuild on compute

[Matt Riedemann]

During a rebuild in nova-compute we get the image once here:

https://github.com/openstack/nova/blob/3b879f49b3969f1d168eabf1ca6076bf6473a4fe/nova/compute/manager.py#L2984

And then again here:

https://github.com/openstack/nova/blob/3b879f49b3969f1d168eabf1ca6076bf6473a4fe/nova/compute/manager.py#L3131

Which is wasteful since we can just pass the already-retrieved image_meta from the first method to the second.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/600260

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/600260
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=0c33cbb81f4c74486f101f4f10afd4a5e1050504
Submitter: Zuul
Branch: master

commit 0c33cbb81f4c74486f101f4f10afd4a5e1050504
Author: Matt Riedemann <email address hidden>
Date: Wed Sep 5 17:34:17 2018 -0400

    Remove redundant image GET call in _do_rebuild_instance

    During a rebuild on the compute, the rebuild_instance method
    fetches the new image from glance once to get the image_meta
    for the RT.rebuild_claim, and then _do_rebuild_instance gets
    the new image again to get image_meta for the driver.spawn/rebuild
    method. Since it's the same image_meta either way, we can just
    pass the image_meta from rebuild_instance to _do_rebuild_instance
    and avoid the redundant GET call to glance.

    Change-Id: I037a83f0a8cb4cdf87514dd4133da8bfedb564f5
    Closes-Bug: #1790972

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/nova 19.0.0.0rc1 release candidate.