1. create a stack enable rollback
2. the stack create failed, and rollback complete, the resources has been deleted
3. stack-show, the stack is in ROLLBACK_COMPLETE
4. to list resources of the stack, some information of resource such as updated_time, creation_time and status are incorrect, the result like this:
root@monasca:/usr/hth# heat resource-list feee80bc-67b2-4582-99ce-c008f34b35c2
+---------------+----------------------+--------------------+-----------------+----------------------------+
| resource_name | physical_resource_id | resource_type | resource_status | updated_time |
+---------------+----------------------+--------------------+-----------------+----------------------------+
| cv | | OS::Cinder::Volume | INIT_COMPLETE | 2015-07-31T09:03:59.041671 |
+---------------+----------------------+--------------------+-----------------+----------------------------+
5. The resource's state is wrong, should be DELETE_COMPLETE or ROLLBACK_COMPLETE, and the creation_time and updated_time should not be utcnow; Even should not return the resources which have been deleted.
I met this bug too, and I've analyzed the cause of this bug preliminarily.
The resource records in DB will be deleted when stack rollbacks.
So when listing resources of rollbacking stack, some resources will just have initialized properties which are initialized by resource.__init__, such as status, updated_time, because can't find resource infos in DB.
The code referred to as follow, which is at the end of resource.__init__:
resource = stack.db_ resource_ get(name)
self._ load_data( resource)
if resource:
Hope this can provide some help.