xenapi: failed snapshots never deleted

[John Garbutt]

When uploading a snapshot, its possible that nova-compute process is killed.

When this happens, currently, that snapshot is never deleted, and the VDI chain can grown a lot.

To fix this, we should remove any snapshots from the chain, before taking the next snapshot.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/100880

[OpenStack Infra (hudson-openstack)]

Change abandoned by John Garbutt (<email address hidden>) on branch: master
Review: https://review.openstack.org/100880
Reason: Agreed direction now, getting rid of this one.

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/100880
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=64520ceeacdb42b66d3b15474124d273d9e8ccf2
Submitter: Jenkins
Branch: master

commit 64520ceeacdb42b66d3b15474124d273d9e8ccf2
Author: John Garbutt <email address hidden>
Date: Wed Jun 18 13:19:57 2014 +0100

    XenAPI: Remove interrupted snapshots

    Currently the VDI chain can grow very long when a snapshots happen at
    the same time as nova-compute being terminated.

    While we now clean up the instance state, the VDI chain is left in a bad
    state, it has an extra snapshot that is no longer required.

    This change improves that by looking at when we detect a failed
    snapshot, we go back and tidy up the VDI chain.

    Partial-Bug: #1331440

    Change-Id: I9bae82048910d8c45bc2a4093064c1ac68f15750