for osd in sorted(self._get_osd_df()['nodes'], key=lambda x: sorted(x.keys())):