I have built a cloudera cluster which contains hdfs, yarn ,spark, zooker, hbase, hive, oozie and hue. Some system disk(/dev/sda) error occurs in one machine. We need to repair this disk and reinstall linux. After Linux installed. Then I reinstall cloudera-agent but find that all cloudera-agent configure file are lost. And our hdfs, yarn dictionary are still in other disk. How do I resume this machine to cluster?