mgmt-reset-server-status sets trove server status to error

If for some reason trove instance remains in an intermediate state such as building, you can use this management capability to manually reset the server status to error state. You can then delete the instance.

ex:

if a instance is stuck in status=BUILD, the database likely resembles

compute_instance_id: d3699fa8-89aa-4768-847a-8ebc11ac94c8
             task_id: 1
    task_description: No tasks for the instance.
     task_start_time: NULL
           volume_id: fb4ccf7f-6918-439b-8e02-f59edbff5960
       server_status: NULL
             deleted: 0
          deleted_at: NULL

mysql> select * from service_statuses where instance_id='1e72b8fb-8a18-4a75-8950-b46470c818ab';
+--------------------------------------+--------------------------------------+-----------+--------------------+---------------------+
| id | instance_id | status_id | status_description | updated_at |
+--------------------------------------+--------------------------------------+-----------+--------------------+---------------------+
| a2c511c0-101c-4c90-aa84-bc98c4cdb54e | 1e72b8fb-8a18-4a75-8950-b46470c818ab | 23 | new | 2014-05-15 00:38:41 |
+--------------------------------------+--------------------------------------+-----------+--------------------+---------------------+

and in this state, mgmt-reset-task-status accomplishes nothing and neither admin or the user can delete (a delete fails with 'instance isn't ready')