Fix db restore reliability if replacing existing

If there is an existing DB, there were race conditions in loading the
restore.  Fix by clearing the existing DB once the viability of the
restore dataset has been established

confluent_server/confluent/config/configmanager.py

@@ -1347,7 +1347,9 @@ class ConfigManager(object):
                     pass
         # Now we have to iterate through each fixed up element, using the
         # set attribute to flesh out inheritence and expressions
+        _cfgstore['main']['idmap'] = {}
         for confarea in _config_areas:
+            self._cfgstore[confarea] = {}
             if confarea not in tmpconfig:
                 continue
             if confarea == 'nodes':