Let's get started with Hadoop using the following steps:
- Format NameNode.
Before starting up the Hadoop daemons, we need to format NameNode. This is like formatting a new disk. We know that just as we bring a new disk drive and format it to a file system (FAT, NTFS, ExtFS), we format NameNode. This process creates a logical layer of HDFS on top of the file system.
Note
<hadoop directory path>can be added to the system path to call thehadoopcommand directly. Follow http://www.cyberciti.biz/faq/unix-linux-adding-path/ for various options.The following line will format NameNode and make it available to read and write files from HDFS. This command prepares the HDFS file system:
<hadoop directory path>/bin/hdfsnamenode –format <cluster name> - Once NameNode is formatted successfully without any errors, let's start the Hadoop daemons:
sbin/hadoop-daemon.sh start namenode sbin/hadoop-daemons.sh start datanode sbin/yarn-daemon.sh start resourcemanager sbin/yarn-daemons.sh start nodemanager sbin/yarn-daemon.sh start historyserver
- Once all the processes start, we can check it using the following command:
ps –ef|grep JavaUse the following command found in
<Java home>/bin/jps:Jps - For easy access of
jps, we can create an alias and just typejps. It will display the Java process running;jpsstands for Java process. As all Hadoop daemons are Java processes,jpsdisplays the daemons running:alias jps=<Java path>/bin/jpsNow, we can directly access the
jpscommand. - To stop the Hadoop cluster, we can use:
sbin/hadoop-daemon.sh stop namenode sbin/hadoop-daemons.sh stop datanode sbin/yarn-daemon.sh stop resourcemanager sbin/yarn-daemons.sh stop nodemanager sbin/yarn-daemon.sh stop historyserver
Now, let's move forward and configure HBase.