slapgrid: Execute manager even if promise fails

Despite promise failure, managers shall be executed, as often thanks to
running them, promise can be fulfilled.

slapos/grid/slapgrid.py

@@ -1247,8 +1247,8 @@ stderr_logfile_backups=1
     finally:
       self.logger.removeHandler(partition_file_handler)
       partition_file_handler.close()
-
-    # Run manager tear down
+      # Run manager tear down, even if something happened, like promise error,
+      # as manager might be used for this
       for manager in self._manager_list:
         manager.instanceTearDown(local_partition)