Sorry, you need to enable JavaScript to visit this website.

Thanks to Mohit Singh for developing a step by step guide: 

Checked on Ubuntu: 12.10 64-bit pseudo-distributed mode
Prerequisites:
  1. Hadoop node setup (Refer: http://www.michael-noll.com/tutorials/running-hadoop-on-ubuntu-linux-sin...)
  2. Maven install (can be set-up thru sudo apt-get install maven on debian machines)
             
For this setup: hadoop was installed at /usr/local/hadoop and Hadoop user account: hduser 
 
Step1: Login to your local hadoop account and download (or checkout) graphbuilder version to your desired location and extract the files.
            %hduser@hadoop: cd /home/hduser
                                     (or git clone https://github.com/01org/graphbuilder.git)
            %hduser@hadoop: tar -xvfz graphbuilder-beta.tar 
 
Step2: Compile the source using:
            %hduser@hadoop:/home/hduse/grapbuilder-beta/grapbuilder$ mvn package
            (A jar file will be created at graphbuilder/target graphbuilder-0.0.1-SNAPSHOT-hadoop-job.jar)
 
Step3: Download a sample wiki file  and extract the compressed file:
            %hduser@hadoop: tar -xvfz enwiki-latest-pages-articles1.xml-p000000010p000010000.bz2
 
Step4: Copy the files to hdfs:
            %hduser@hadoop: $cd/usr/local/hadoop
            %hduser@hadoop:/usr/local/hadoop$ bin/hadoop dfs -copyFromLocal /home/hduser/enwiki-latest-pages-articles1.xml-p000000010p000010000 /user/hduser/wiki-input
            Check the dfs (sanity check)
            %hduser@hadoop:/user/local/hadoop$ bin/hadoop dfs -ls
            Found 1 item
            -rw-r--r--   1 hduser supergroup  148357153 2013-01-02 12:21 /user/hduser/wiki-input
 
Step5: To run a demoapp (for example: linkgraph example)
            %hduser@hadoop: bin/hadoop jar /home/hduser/graphbuilder/target/graphbuilder-0.0.1-SNAPSHOT-hadoop-job.jar com.intel.hadoop.graphbuilder.demoapps.wikipedia.linkgraph.LinkGraphEnd2End 3 /user/hduser/en-wiki-articles /user/hduser/en-wiki-articles-output ingressCode*
     
            This should start the hadoop jobs. After jobs has been succesfully executed.
            For sanity check:
            %hduser@hadoop:/user/local/hadoop bin/hadoop dfs -ls
            Found 2 items:
            -rw-r--r--   1 hduser supergroup  148357153 2013-01-02 12:21 /user/hduser/wiki-input
            -rw-r--r--   1 hduser supergroup  148357153 2013-01-02 12:21 /user/hduser/wiki-articles-output
 
  * ingressCode specifies the graph partitioning algorithm. Valid options are "0, 1, 2, 3, 4, 5, 6 and 7". Please refer to the white paper or GraphBuilder paper for further information.
 
Known Issues:
 It may so happen that some of the reducer jobs might take longer time and you may see the error message
    13/01/02 13:19:42 INFO mapred.JobClient: Task Id : attempt_201301021121_0010_m_000000_0, Status : FAILED
    Task attempt_201301021121_0010_m_000000_0 failed to report status for 600 seconds. Killing!
 
You might have to add the following parameters in your hadoop/conf/mapred-site.xml
<property>
   <name>mapred.task.timeout</name>
   <value>1800000</value> 
</property>
 
Project: