Time for action – killing the NameNode process

Let's now kill the NameNode process, which we should expect to directly stop us from accessing HDFS and by extension, prevent the MapReduce jobs from executing:

Note

Don't try this on an operationally important cluster. Though the impact will be short-lived, it effectively kills the entire cluster for a period of time.

1. Log onto the NameNode host and list the running processes:

   $ jps
   2372 SecondaryNameNode
   2118 NameNode
   2434 JobTracker
   5153 Jps
   

2. Kill the NameNode process. Don't worry about SecondaryNameNode, it can keep running.

3. Try to access the HDFS filesystem:

   $ hadoop fs -ls /
   11/12/13 16:00:05 INFO ipc.Client: Retrying connect to server: /10.0.0.100:9000. Already tried 0 time(s).
   11/12/13 16:00:06 INFO ipc.Client: Retrying connect to server: /10.0.0.100:9000. Already tried 1 time(s).
   11/12/13 16:00:07 INFO ipc.Client: Retrying connect to server: /10.0.0.100:9000. Already tried 2 time(s).
   11/12/13 16:00:08 INFO ipc.Client: Retrying connect to...