Snapshots aren't cleaned up when coalesce fails

[Rick Harris]

If two VHDs aren't able to coalesce into each other, they will eventually timeout and raise an exception. In vmops:_create_snapshot this exception is swallowed preventing the snapshot from getting torn down.

We modify the code to re-raise the exception so we can cleanup.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/5058

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/5058
Committed: http://github.com/openstack/nova/commit/4d7f428a2809765ed1c5e05ad790b1e33d905801
Submitter: Jenkins
Branch: master

commit 4d7f428a2809765ed1c5e05ad790b1e33d905801
Author: Rick Harris <email address hidden>
Date: Thu Mar 8 01:45:39 2012 +0000

    Don't swallow snapshot exceptions.

    Fixes bug 949471

    This raise is needed so that calling code can be notified to teardown
    the failed snapshot.

    Change-Id: I25a78e36b4bdb4ca52f1f999168a750dde134056