By | August 10, 2015

Hadoop installation on windows without cygwin in 10 mints – Hadoop installation on windows 7 or 8

Before starting make sure you have this two softwares

Extract downloaded tar file

Step 1 – Windows path configuration
set HADOOP_HOME path in enviornment variable for windows
Right click on my computer > properties > advanced system settings > advance tab > environment variables > click on new

Set hadoop bin directory path
Find path variable in system variable > click on edit > at the end insert ‘; (semicolon)’ and paste path upto hadoop bin directory in my case it’s a


Step 2 – Hadoop configuration
Edit hadoop-2.7.1/etc/hadoop/core-site.xml, paste the following lines and save it.


Edit hadoop-2.7.1/etc/hadoop/mapred-site.xml, paste the following lines and save it.


Edit hadoop-2.7.1/etc/hadoop/hdfs-site.xml, paste the following lines and save it, please create data folder somewhere and in my case i have created it in my HADOOP_HOME directory


Edit hadoop-2.7.1/etc/hadoop/yarn-site.xml, paste the following lines and save it.


Edit hadoop-2.7.1/etc/hadoop/hadoop-env.cmd, comment existing %JAVA_HOME% using @rem at start, give proper path and save it. (my jdk is in program files to avoid spaces i gave PROGRA~1)


Step 3 – Start everything

Very Important step

Before starting everything you need to add some [dot].dll and [dot].exe files of windows please download bin folder from my github repository – sardetushar_gitrepo_download

bin folder – this contains .dll and .exe file (winutils.exe for hadoop 2.7.1)

Now delete you existing bin folder and replace with new one (downloaded from my repo )

( my github ) etc folder is just given for reference you need to modify your configuration parameters according to your environment path

Open cmd and type ‘hdfs namenode -format’ – after execution you will see below logs

Open cmd and point to sbin directory and type ‘start-all.cmd


It will start following process



YARN resourcemanager

YARN nodemanager

JPS – to see services are running
open cmd and type – jps (for jps make sure your java path is set properly)

Step 4 – namenode GUI, resourcemanager GUI
Resourcemanager GUI address – http://localhost:8088

Namenode GUI address – http://localhost:50070

In next tutorial we will see how to run mapreduce programs in windows using eclipse and this hadoop setup

Share this knowledge ! Join us on Facebook ! Now Whatsapp sharing is supportable ! TooDey Inc.

2 thoughts on “

  1. Juan

    Can someone help me with this error,
    DEPRECATED: Use of this script to execute hdfs command is deprecated.
    Instead use the hdfs command for it.
    Error: Could not find or load main class J

    Where I can find the J class?

    Thanks for your respond 🙂

  2. Alexey

    Hi! It was really useful for me, thanks a lot! Can we also install hbase same way?


Leave a Reply

Your email address will not be published. Required fields are marked *