Hadoop Web Interfaces

Hadoop provides several web interfaces to monitor your Hadoop cluster. These web interfaces give a lot of information without dealing with shell commands.

I am using a Hadoop 1.0.3 cluster installed on Centos machines.

1. NameNode web interface

http://<namenode-server>:50070/

• Enables user to browse HDFS file system
• Gives summary about the cluster
• This is the default address and port. Address and port can be changed by dfs.http.address property in hdfs-site.xml

2. JobTracker web interface

http://<jobtracker-server>:50030/

• Gives information about the tasktracker nodes and mapreduce capacity of Hadoop cluster
• Lists running/completed/retired jobs
• Enables user to browse job history
• Address and port can be changed by mapred.job.tracker.http.address property in mapred-site.xml

3. TaskTracker web interface

http://<tasktracker-server>:50060/

• List running and non-running tasks on specific datanode
• Address and port can be changed by mapred.task.tracker.http.address property in mapred-site.xml

4. Secondary NameNode web interface

http://<secondaryNN-server>:50090/

Address and port can be changed by dfs.secondary.http.address property in hdfs-site.xml

5. DataNode web interface

http://<datanode-server>:50075/

• Enables browsing of data on specific datanode
• Address and port can be changed by dfs.datanode.http.address property in hdfs-site.xml