Assert volume deletion in test_image_defined_boot_from_volume

Recently, we had a regression in the cinder api which resulted in the
inability to detach (and thus delete) a volume that had been created
from an image. This adds an assertion that the underlying volume has
been deleted in order to catch future regressions.

Related-Bug: #1768650

Change-Id: I5ef43b11e01f3c8980dcb364b23edda1b7af2859
diff --git a/tempest/scenario/test_volume_boot_pattern.py b/tempest/scenario/test_volume_boot_pattern.py
index 1fc57e7..2d024e9 100644
--- a/tempest/scenario/test_volume_boot_pattern.py
+++ b/tempest/scenario/test_volume_boot_pattern.py
@@ -236,16 +236,11 @@
 
         # Delete the second server which should also delete the second volume
         # created from the volume snapshot.
-        # TODO(mriedem): Currently, the compute service fails to delete the
-        # volume it created because the volume still has the snapshot
-        # associated with it, and the cleanups for the volume snapshot which
-        # were added in create_server_snapshot above, don't run until after
-        # this is called. So we need to delete the volume snapshot and wait for
-        # it to be gone here first before deleting the server, and then we can
-        # also assert that the underlying volume is deleted when the server is
-        # deleted.
         self._delete_server(instance)
 
+        # Assert that the underlying volume is gone.
+        self.volumes_client.wait_for_resource_deletion(created_volume['id'])
+
     @decorators.idempotent_id('cb78919a-e553-4bab-b73b-10cf4d2eb125')
     @testtools.skipUnless(CONF.compute_feature_enabled.attach_encrypted_volume,
                           'Encrypted volume attach is not supported')