Fix cluster_application_kill method

The recently introduced cluster_application_kill() method was erroneously
calling `self.put()` to issue the status change to `KILLED`.  This resulted
in an AttributeError exception:
```
''ResourceManager' object has no attribute 'put''
```
in client code.

The base class issues put calls, via its `update()` method.  Once updated,
application kill requests succeed.

Author: kevin-bates

yarn_api_client/resource_manager.py

@@ -213,7 +213,7 @@ def cluster_application_kill(self, application_id):
         path = '/ws/v1/cluster/apps/{appid}/state'.format(
             appid=application_id)
 
-        return self.put(path, data)
+        return self.update(path, data)
 
     def cluster_nodes(self, state=None, healthy=None):
         """