Amazon EMR
Developer Guide

Viewing and Restarting Amazon EMR and Application Processes (Daemons)

This documentation is for AMI versions 2.x and 3.x of Amazon EMR. For information about Amazon EMR releases 4.0.0 and above, see the Amazon EMR Release Guide. For information about managing the Amazon EMR service in 4.x releases, see the Amazon EMR Management Guide.

When you troubleshoot a cluster, you may want to list running processes. You may also find it useful to stop or restart processes in some circumstances—for example, after you change a configuration or notice a problem with a particular process after you analyze log files and error messages.

There are two types of processes that run on a cluster: Amazon EMR processes (for example, instance-controller and Log Pusher), and processes associated with the applications installed on the cluster (for example, hadoop-hdfs-namenode, and hadoop-yarn-resourcemanager).

To work with processes directly on a cluster, you connect to the master node. For more information, see Connect to the Cluster.

Viewing Running Processes

If you are using Amazon EMR versions earlier than 4.x, both Amazon EMR processes and processes associated with applications you install run as part of the SysV init system via init.d scripts.

To view a list of running processes

  • Type the following command (without the $, which indicates the Linux command prompt):

    $ ls /etc/init.d/

    The command returns a list of running processes similar to the following example:

    acpid hadoop-hdfs-namenode instance-controller

Restarting Processes

After you determine which processes are running, you can stop and then restart them if necessary.

To restart a process

  1. Type the following command to stop the process, replacing processname with the process name returned by the ls command in the procedure above:

    $ sudo /etc/init.d/processname stop

    For example: sudo /etc/init.d/hadoop-hdfs-namenode stop

  2. Type the following command to restart the process:

    $ sudo /etc/init.d/processname start

    For example, sudo /etc/init.d/hadoop-hdfs-namenode start.