@Koffi This issue seems linked to your previous posting. Your last healthy name node was nn01, right? The assumption here is you are logged in as root Instructions to fix that one journal node. 1) Put both nn01 and nn02 in safe mode ( NN HA) $ sudo su - hdfs
[hdfs@host ~]$ hdfs dfsadmin -safemode enter Safe mode is ON in nn01/<nn01_IP>:8020 Safe mode is ON in nn02/<nn02_IP>:8020 2) Save Namespace [hdfs@host ~]$ hdfs dfsadmin -saveNamespace Save namespace successful for nn01/<nn01_IP>:8020 Save namespace successful for nn02/<nn02_IP>:8020 3) Backup zip/tar the journal dir from a working JN node of (nn01) and copy it to the non-working JN's of (nn02)node to something like /hadoop/hdfs/journal/<Cluster_name>/current 4) Leave safe mode [hdfs@host ~]$ hdfs dfsadmin -safemode leave Safe mode is OFF in nn01/<nn01_IP>:8020 Safe mode is OFF in nn02/<nn02_IP>:8020 4) Restart HDFS From Ambari you can now start the nn01 first when it comes up then start nn02 Please let me know.
... View more
