xenapi unrescue sometimes leaves rescue VM undeleted

[Johannes Erdfelt]

There's a race condition between the shutdown and the destroy which results in the rescue VM sometimes not being destroyed.

This is caused by the use of an asynchronous shutdown but without waiting for it to finish.

Since the rescue VM is not destroyed any subsequent attempts to rescue an instance will fail.

Also, the destroy is done asynchronously without any wait so any failures will be silently dropped.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/5094

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/5094
Committed: http://github.com/openstack/nova/commit/5555446d877223341c2723e4be169d996ab2ce3b
Submitter: Jenkins
Branch: master

commit 5555446d877223341c2723e4be169d996ab2ce3b
Author: Johannes Erdfelt <email address hidden>
Date: Thu Mar 8 19:13:26 2012 +0000

    Wait for rescue VM shutdown to complete before destroying it

    Fixes bug 950168

    Also, wait for the VM and VDI destroys to complete so any errors can
    get logged

    Change-Id: I2c2237ed6f95828198824db8628d267ac5b26b3a