Hadoop is an open-source framework that is widely used to deal with Bigdata. Most of the Bigdata/Data Analytics projects are being built up on top of the Hadoop Eco-System. It consists of two-layer, one is for Storing Data and another one is for Processing Data.

Storage will be taken care of by its own filesystem called HDFS (Hadoop Distributed Filesystem) and Processing will be taken care of by YARN (Yet Another Resource Negotiator). Mapreduce is the default processing engine of the Hadoop Eco-System.



  • A minimal installation of CentOS 7 server.
  • Java v1.8 release.
  • Hadoop 2.x stable release.


Let have a check with steps and process involved to set up Hadoop Pseudonode (Single NodeCluster


Installing Java on CentOS 7


1. Hadoop is an Eco-System which is made up of Java. We need Java installed in our system mandatorily to install Hadoop.

# yum install java-1.8.0-openjdk


2. Next, verify the installed version of Java on the system.

# java -version


Configure Passwordless Login on CentOS 7


We need to have ssh configured in our machine, Hadoop will manage nodes with the use of SSH. Master node uses SSH connection to connect its slave nodes and perform operation like start and stop.


We need to set up password-less ssh so that the master can communicate with slaves using ssh without a password. Otherwise for each connection establishment, need to enter the password. 


In this single node, Master services (Namenode, Secondary Namenode & Resource Manager) and Slave services (Datanode & Nodemanager) will be running as separate JVMs. Even though it is single node, we need to have password-less ssh to make Master to communicate Slave without authentication.


3. Set up a password-less SSH login using the following commands on the server.

# ssh-keygen
# ssh-copy-id -i localhost


4. After you configured passwordless SSH login, try to login again, you will be connected without a password.

# ssh localhost


Installing Hadoop in CentOS 7


5. Go to the Apache Hadoop website and download the stable release of Hadoop using the following wget command.

# wget https://archive.apache.org/dist/hadoop/core/hadoop-2.10.1/hadoop-2.10.1.tar.gz
# tar xvpzf hadoop-2.10.1.tar.gz


6. Next, add the Hadoop environment variables in ~/.bashrc file as shown.



7. After adding environment variables to ~/.bashrc the file, source the file and verify the Hadoop by running the following commands.

# source ~/.bashrc
# bin/hadoop version


Configuring Hadoop in CentOS 7

We need to configure below Hadoop configuration files in order to fit into your machine. In Hadoop, each service has its own port number and its own directory to store the data.

  • Hadoop Configuration Files – core-site.xml, hdfs-site.xml, mapred-site.xml & yarn-site.xml


8. First, we need to update JAVA_HOME and Hadoop path in the hadoop-env.sh file as shown.

# cd $HADOOP_PREFIX/etc/hadoop
# vi hadoop-env.sh


Enter the following line at beginning of the file.

export JAVA_HOME=/usr/lib/jvm/java-1.8.0/jre
export HADOOP_PREFIX=/root/hadoop-2.10.1


9. Next, modify the core-site.xml file.

# cd $HADOOP_PREFIX/etc/hadoop
# vi core-site.xml

Paste following between <configuration> tags as shown.



10. Create the below directories under rootadminz user home directory, which will be used for NN and DN storage.

# mkdir -p /home/rootadminz/hdata/
# mkdir -p /home/rootadminz/hdata/data
# mkdir -p /home/rootadminz/hdata/name


11. Next, modify the hdfs-site.xml file.

 # cd $HADOOP_PREFIX/etc/hadoop # vi hdfs-site.xml/p>


Paste following between <configuration> tags as shown.

          dfs .datanode.data.dir


12. Again, modify the mapred-site.xml file.

# cd $HADOOP_PREFIX/etc/hadoop
# cp mapred-site.xml.template mapred-site.xml
# vi mapred-site.xml


Paste following between <configuration> tags as shown.



13. Lastly, modify the yarn-site.xml file.

# cd $HADOOP_PREFIX/etc/hadoop
# vi yarn-site.xml


Paste following between <configuration> tags as shown.



Formatting the HDFS File System via the NameNode


14. Before starting the Cluster, we need to format the Hadoop NN in our local system where it has been installed. Usually, it will be done in the initial stage before starting the cluster the first time.

Formatting the NN will cause loss of data in NN metastore, so we have to be more cautious, we should not format NN while the cluster is running unless it is required intentionally.

# bin/hadoop namenode -format


15. Start NameNode daemon and DataNode daemon: (port 50070).

# sbin/start-dfs.sh


16. Start ResourceManager daemon and NodeManager daemon: (port 8088).

# sbin/start-yarn.sh


17. To stop all the services.

# sbin/stop-dfs.sh
# sbin/stop-dfs.sh



Дали Ви помогна овој одговор? 0 Корисниците го најдоа ова како корисно (0 Гласови)