Setup Hadoop YARN on CentOS VMs

This post summarizes my experience in setting up testing environment for YARN using CentOS  VMs.

After we setup the hdfs following the link (http://czcodezone.blogspot.sg/2016/03/setup-hdfs-cluster-in-centos-vms.html). We can go ahead to set up yarn to manage jobs in hadoop.

Hadoop v2 uses application masters in the datanode to work together with nodemanager to manage a job, and uses resource manager in the namenode to schedule resources for a job (a job refers to a particular application or driver from distributed computation framework such as mapreduce or spark).

1. Setup yarn configuration in hadoop

To setup yarn, on each VM (both namenode and datanode), perform the following steps:

1.1. Edit hadoop/etc/hadoop/mapred-site.xml

Run the following command to edit the mapred-site.xml:

```bash
`cd hadoop/etc/hadoop
`cp mapred-site.xml.template mapred-site.xml
`vi mapred-site.xml

In the mapred-site.xml, Modify as follows:

<configuration>
<property>
<name>fs.default.name</name>
<value>hdfs://centos01:9000</value>
</property>
</configuration>

The "hdfs://centos01:9000" specify the master as the "fs.default.name". (It is important that this is not specified as "hdfs://localhost:9000", otherwise the "hadoop/bin/hdfs dfsadmin -report" will have "Connection refused" exception)

1.2 Edit hadoop/etc/hadoop/yarn-site.xml

Run the following command to edit the yarn-site.xml:

```bash
`vi hadoop/etc/hadoop/yarn-site.xml

In the yarn-site.xml, modify as follows:

<configuration>
<property>
<name>yarn.nodemanager.aux-services</name>
<value>mapreduce_shuffle</value>
</property>
<property>
<name>yarn.nodemanager.aux-services.mapreduce_shuffle.class</name>
<value>org.apache.hadoop.mapred.ShuffleHandler</value>
</property>
</configuration>

2. Start the yarn in the namenode.

On the namenode centos01 (please refers to http://czcodezone.blogspot.sg/2016/03/setup-centos-vm-in-virtualbox-for.html), run the following command to start the hdfs and then yarn:

```bash
`hadoop/sbin/start-dfs.sh
`hadoop/sbin/start-yarn.sh
`hadoop/sbin/mr-jobhistory-daemon.sh start historyserver
`jps

3. Stop hdfs and yarn

To stop hdfs and yarn, run the following command:

```bash
`hadoop/sbin/mr-jobhistory-daemon.sh stop historyserver
`hadoop/sbin/stop-yarn.sh
`hadoop/sbin/stop-dfs.sh