Setting Up Hadoop Made Easy

Become a Certified Hadoop Developer on udemy by Nitesh Jain.
Look for Become a Certified Hadoop Developer on www.udemy.com
Setting up Hadoop made Easy

Word of Motivation
Hadoop installation is the most complex step when you start out to learn Hadoop, especially when you
are new to Linux as well. At some point of time it may test you, please be patient and follow the steps
below. Many have installed it following the same steps as below.
Although I have tried to cover installation which should be applicable to all scenarios, but some strange
situation specific error can spring up at your end. When Hadoop tests you with a challenge, please try
to resolve it through internet.
Just in case, if you fail to get the right advice on internet and are stuck for long (2 days or more),
please contact me. I would help you out.
Basic Idea in a Nutshell

Following are the steps that would be taken in a nutshell:
1.
2.
3.
4.
5.
6.
7.
Install virtual machine on windows or OS.

Install Ubuntu on the virtual machine.
Download and untar Hadoop package on Ubuntu.
Download and install Java on Ubuntu. (Hadoop is written completely in Java).
Tell Ubuntu where the Java installation has been done.
Tell Hadoop where Java installation has been done. At this point Standalone is done.
For pseudo-distribution mode, change the configuration files to configure:
a. Core-site.xml -> to set default Schema and authority.
b. Hdfs-site.xml -> to set def.replication to 1 rather than the default three, otherwise all
the blocks would always be alarmed with under replication.
c. Mapred-site.xml -> To let know of host and port pair where the Jobtrackers runs at.
8. Format the name node and you are ready.
Version details
Following are the details of components used, all license free:
1. Hadoop 1.2.1
2. Ubuntu LTS 12.04 (running on virtual Machine) 64 Bit
3. Windows 8. (The same thing can be done on mac, i.e., install a virtual machine on mac and
follow the below procedure). Any windows machine would do well.
Step 1. Installing Virtual Machine

Step 1.1 Download
Free version of Oracle VirtualBox can be downloaded from:

https://www.virtualbox.org/wiki/Downloads
Download UBUNTU LTS 64 bit from the following link (Make sure its ISO format and for 64 bit):
http://www.ubuntu.com/download/desktop
Step 1.2 Installation








In the below screen shot click on the + sign to add ISO

which you have already downloaded to be loaded as CD
drive.
Press Start.

If throws an error, saying something about that 64 bit support and about
VT-x/AMD-V,
It means that your BIOS doesnt support visualization.

Perform the following steps. This is for my configuration yours may be a little different:
o Restart you computer and go to BIOS setup
o Goto UEFI Firmware>>Advaced>>CPU Setup >> Intel Virtualization Techonlogy.
Enable this.
o Save and exit.
Now try to start the Ubuntu boot with the ISO image and it should work.


Click on install Ubuntu.


And after you have pressed continue the whole disk would
be formatted!
Nope just joking! (: Only the dynamic Disk allocated would be formatted.


(I live in Melbourne. One of the loveliest cities in the world.)


Step 2 Download Hadoop tar.gz

At this point you would like to reopen this document on Ubuntu. You can transfer it by internet.
(Most of the following steps are referred from apache docs:
http://hadoop.apache.org/docs/stable/single_node_setup.html
The link above might get obsolete with time. If so, please google search Apache hadoop installation to
find apache installation guide)
The main idea behind the following steps is to create a folder for hadoop and untar (or unzip) the tar
file that has been downloaded.
1.
2.
3.
4.
5.
Downloading a stable release copy ending with tar.gz

Create a new folder /home/hadoop
Move the file hadoop.x.y.z.tar.gz to the folder /home/Hadoop
Type/Copy/Paste: cd /home/hadoop
Type/Copy/Paste: tar xzf hadoop*tar.gz
Step 3 Downloading and setting up Java

For more refer: http://www.wikihow.com/Install-Oracle-Java-on-Ubuntu-Linux
(The link above might get obsolete with time. If so, please google search wikihow install java Ubuntu
Linux)

1. Check if Java is already present, by

Type/Copy/Paste : java version
2. If it is 1.7.* then you can setup the JAVA_HOME Variable according to where it is setup.
3. If you are confident to setup the JAVA_HOME variable please go ahead to step 9 (in this section
itself). If not dont worry and follow the following steps:
4. First we will purge the Java installed.
Type/Copy/Paste : sudo apt-get purge openjdk-\*
5. Make the directory where java would installed, by:
sudo mkdir -p /usr/local/java
6. Download Java JDK and JRE from the link, look for linux, 64 bit and tar.gz ending file:
http://www.oracle.com/technetwork/java/javase/downloads/index.html
7. Goto downloads folder and then copy to the folder we created for java:
Type/Copy/Paste: sudo cp -r jdk-*.tar.gz /usr/local/java
Type/Copy/Paste: sudo cp -r jre-*.tar.gz /usr/local/java
8. Extract and install Java:
Type/Copy/Paste: cd /usr/local/java
Type/Copy/Paste: sudo tar xvzf jdk*.tar.gz
Type/Copy/Paste: sudo tar xvzf jre*.tar.gz
9. Now put all the variables in the profile.
Type/Copy/Paste: sudo gedit /etc/profile
At the end copy paste the following.(Note: change the highlighted paths according to your
installations. Version number would have changed from making this guide to your installation.
So just make sure that the path you mention actually exists)
[Tip: It is important that you specify complete and correct path while declaring HADOOP_INSTALL
variable. To get the right value, navigate to the folder hadoop-1.2.1 and then type in command
pwd which would return the complete present working directory. Copy and paste that to avoid
any typo errors.]
JAVA_HOME=/usr/local/java/jdk1.7.0_40
PATH=$PATH:$JAVA_HOME/bin
JRE_HOME=/usr/local/java/jre1.7.0_40
PATH=$PATH:$JRE_HOME/bin
HADOOP_INSTALL=/home/{user_name}/hadoop/hadoop-1.2.1
PATH=$PATH:$HADOOP_INSTALL/bin
export JAVA_HOME
export JRE_HOME
export PATH
10. Do the following so that Linux knows where Java is, (Note that the highlighted following paths
may be needed to changed in accordance to your installation):
sudo update-alternatives --install "/usr/bin/java" "java" "/usr/local/java/jre1.7.0_40/bin/java" 1

sudo update-alternatives --install "/usr/bin/javac" "javac" "/usr/local/java/jdk1.7.0_40/bin/javac" 1
sudo update-alternatives --install "/usr/bin/javaws" "javaws" "/usr/local/java/jre1.7.0_40/bin/javaws" 1
sudo update-alternatives --set java /usr/local/java/jre1.7.0_40/bin/java
sudo update-alternatives --set javac /usr/local/java/jdk1.7.0_40/bin/javac
sudo update-alternatives --set javaws /usr/local/java/jre1.7.0_40/bin/javaws
11. Refresh the profile by:

Type/Copy/Paste: . /etc/profile
12. Test by typing Java version.
Step 4 Stand Alone mode installed! Congratulations!

At this point you should have had got to the point that you can run Hadoop in Stand Alone mode. You
can practice almost anything for practicing developments in Map Reduce. Test if you are successful:
Type/Copy/Paste: cd /home/hadoop
(going to the Hadoop directory)
Type/copy/Paste: mkdir input
Type/copy/Paste: bin/hadoop jar hadoop-examples-*.jar grep input output 'dfs[a-z.]+'
Or the above can be typed in without bin as well.
Type/copy/Paste: hadoop jar hadoop-examples-*.jar grep input output 'dfs[a-z.]+'
Type/copy/Paste: ls output/*
Step 5 Pseudo Distribution Mode

1. Type/Copy/Paste: sudo apt-get install ssh
2. Type/Copy/Paste: sudo apt-get install rsync
3. Change conf/core-site.xml to:
(to install ssh)
<configuration>
<property>
<name>fs.default.name</name>
<value>hdfs://localhost:9000</value>
</property>
</configuration>
4. Change conf/hdfs-site.xml to:
<configuration>
<property>
<name>dfs.replication</name>
<value>1</value>
</property>
</configuration>

5. Change conf/mapred-site.xml to:
<configuration>
<property>
<name>mapred.job.tracker</name>
<value>localhost:9001</value>
</property>
</configuration>
6. Edit conf/hadoop-env.sh look for JAVA_HOME and set it up
export JAVA_HOME=/usr/local/java/jdk1.7.0_40
7. Setup passwordless ssh by:

Type/copy/paste: ssh-keygen -t dsa -P '' -f ~/.ssh/id_dsa
Type/copy/paste: cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys
8. To confirm that passwordless ssh has been setup type the following and you shouod not be
prompted for a password.
Type/copy/paste: ssh localhost
9. Format the name node:
Type/copy/paste: bin/hadoop namenode format
10. Start the all the demons:
Type/copy/paste: bin/startall.sh
11. On web browser navigate to http//localhost:50070/ and then to http://localhost:50030/
Make sure hadoop started properly.
http://localhost:50030/ should forward to http://localhost:50030/jobtracker.jsp localhost Hadoop
Map/Reduce Administration page
http://localhost:50070/ should forward to http://localhost:50070/dfshealth.jsp NameNode
'localhost:9000' page
If any of url doesn't work than make sure that namenode and datanode started succussfully by
running the command 'jps' (show java processes) and the output should look like the following:
2310 SecondaryNameNode
1833 NameNode
2068 DataNode
2397 JobTracker
2635 TaskTracker
2723 Jps
If NameNode or DataNode is not listed than it might happen that the namenode's or datanode's root
directory which is set by the property 'dfs.name.dir' is getting messed up. It by default points to
the /tmp directory which operating system changes from time to time. Thus, HDFS when comes up
after some changes by OS, gets confused and namenode doesn't start.
Solution:
a) Stop hadoop by running 'stop-all.sh'
We need to explicitly set the 'dfs.name.dir' and 'dfs.data.dir'.

Perform the following steps and the issue should resolve (You can of course create any
folders and give that path, but below I would be giving an example. You can create your
own folder your way)
b) Goto hadoop folder and create a folder 'dfs'. So now the folder '/home/hadoop/dfs' would
exist. The idea is to make two folders inside it which would be used for datanode demon
and namenode demon.
Create only 'name' folder inside '/home/{user_name}/hadoop/dfs' folder, manually. The
other data folder would be created by hadoop itself by the following steps.
c) Change the configuration file hdfs-site.xml to set properties 'dfs.name.dir' and 'dfs.data.dir'
as follows. Two points to be noted. First, change the indentation. Second, change the
username portion (/{user_name} in the case below, it should be your's) of path according to
your system. Giving incomplete path is a common error:
(TIP: go to newly created dfs folder thorough command prompt and type in command 'pwd'
to get exact path. Copy paste to avoid typos)
configuration file hdfs-site.xml should look like below:
<configuration>
<property>
<name>dfs.data.dir</name>
<value>/home/{user_name}/hadoop/dfs/data/</value>
</property>
<property>
<name>dfs.name.dir</name>
<value>/home/{user_name}/hadoop/dfs/name/</value>
</property>
<property>
<name>dfs.replication</name>
<value>1</value>
</property>
</configuration>
d) Run command
hadoop namenode format
Look for the following output to confirm that the format has been successful. If you do not see
the message, format command is having some problems.
(I am pasting the output of one of the course taker Vadim and so you see the username as
Vadim here)
14/02/04 22:56:12 INFO namenode.NameNode: STARTUP_MSG:
/************************************************************
STARTUP_MSG: Starting NameNode
STARTUP_MSG: host = vadim-VirtualBox/127.0.1.1
STARTUP_MSG: args = [-format]


STARTUP_MSG: version = 1.2.1
STARTUP_MSG: build =https://svn.apache.org/repos/asf/hadoop/common/branches/branch-1.2 -r 1503152;
compiled by 'mattf' on Mon Jul 22 15:23:09 PDT 2013
STARTUP_MSG: java = 1.7.0_51
************************************************************/
Re-format filesystem in /home/vadim/hadoop/dfs/name ? (Y or N) Y
14/02/04 22:56:17 INFO util.GSet: Computing capacity for map BlocksMap
14/02/04 22:56:17 INFO util.GSet: VM type = 64-bit
14/02/04 22:56:17 INFO util.GSet: 2.0% max memory = 1013645312
14/02/04 22:56:17 INFO util.GSet: capacity = 2^21 = 2097152 entries
14/02/04 22:56:17 INFO util.GSet: recommended=2097152, actual=2097152
14/02/04 22:56:18 INFO namenode.FSNamesystem: fsOwner=vadim
14/02/04 22:56:18 INFO namenode.FSNamesystem: supergroup=supergroup
14/02/04 22:56:18 INFO namenode.FSNamesystem: isPermissionEnabled=true
14/02/04 22:56:18 INFO namenode.FSNamesystem: dfs.block.invalidate.limit=100
14/02/04 22:56:18 INFO namenode.FSNamesystem: isAccessTokenEnabled=false accessKeyUpdateInterval=0
min(s), accessTokenLifetime=0 min(s)
14/02/04 22:56:18 INFO namenode.FSEditLog: dfs.namenode.edits.toleration.length = 0
14/02/04 22:56:18 INFO namenode.NameNode: Caching file names occuring more than 10 times
14/02/04 22:56:19 INFO common.Storage: Image file /home/vadim/hadoop/dfs/name/current/fsimage of size 111
bytes saved in 0 seconds.
14/02/04 22:56:19 INFO namenode.FSEditLog: closing edit log: position=4,
editlog=/home/vadim/hadoop/dfs/name/current/edits
14/02/04 22:56:19 INFO namenode.FSEditLog: close success: truncate to 4,
editlog=/home/vadim/hadoop/dfs/name/current/edits
14/02/04 22:56:19 INFO common.Storage: Storage directory /home/vadim/hadoop/dfs/name has been
successfully formatted.
14/02/04 22:56:19 INFO namenode.NameNode: SHUTDOWN_MSG:
/************************************************************
SHUTDOWN_MSG: Shutting down NameNode at vadim-VirtualBox/127.0.1.1
************************************************************/


e) Run command
start-all.sh
f)
Run command
jps
and this will now show all the demons running like the below:
2310 SecondaryNameNode
1833 NameNode
2068 DataNode
2397 JobTracker
2635 TaskTracker
2723 Jps
g) Run command
stop-all.sh
and you should see the output as:


topping jobtracker
localhost: stopping tasktracker
stopping namenode
localhost: stopping datanode
localhost: stopping secondarynamenode
*there should be no message as no namenode to stop or no datanode to stop

Once this is done, you have successfully installed Hadoop in pseudo-distribution mode and this
is one of the most difficult things to do.
If you face any issues, please feel free to post the problem on the question forum so that
everyone can look at it and help.
As well please answer to others problems so that you sharpen your knowledge and others get
the needed help as well.
Best,
Nitesh


Setting Up Hadoop Made Easy

Uploaded by

Copyright:

Available Formats

Setting Up Hadoop Made Easy

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Setting Up Hadoop Made Easy

Uploaded by

Copyright:

Available Formats

Become a Certified Hadoop Developer on udemy by Nitesh Jain.

Look for Become a Certified Hadoop Developer on www.udemy.com