I am using Ubuntu 10.04 LTS – the Lucid Lynx now and for test purpose it is necessary to run the code under different version. More specifically, 0.18 is the stable version for old API; 0.20 is the stable version for new API and 0.21 is the latest test version (0.22 will be the stable version). One of my is related to Distributed Cache, which is quite interesting for different version (Please refer to another post for more details).
To install the three instances of Hadoop is similar to install one. So please read this post first: Install Hadoop
All three instances are located at /usr/local
├── hadoop (for 0.20.2)
HDFS directories layout is:
Note: You can only run one instance even it listens on different ports.
hadoop@fxu-t60:/usr/local/hadoop-0.18.3$ bin/start-all.sh starting namenode, logging to /usr/local/hadoop-0.18.3/bin/../logs/hadoop-hadoop-namenode-fxu-t60.out localhost: starting datanode, logging to /usr/local/hadoop-0.18.3/bin/../logs/hadoop-hadoop-datanode-fxu-t60.out localhost: starting secondarynamenode, logging to /usr/local/hadoop-0.18.3/bin/../logs/hadoop-hadoop-secondarynamenode-fxu-t60.out starting jobtracker, logging to /usr/local/hadoop-0.18.3/bin/../logs/hadoop-hadoop-jobtracker-fxu-t60.out localhost: starting tasktracker, logging to /usr/local/hadoop-0.18.3/bin/../logs/hadoop-hadoop-tasktracker-fxu-t60.out hadoop@fxu-t60:/usr/local/hadoop-0.18.3$ jps 7873 NameNode 8294 JobTracker 8541 Jps 8472 TaskTracker 4911 org.eclipse.equinox.launcher_1.0.201.R35x_v20090715.jar 8051 DataNode 8222 SecondaryNameNode 5261 hadoop@fxu-t60:/usr/local/hadoop$ bin/start-all.sh namenode running as process 7873. Stop it first. localhost: datanode running as process 8051. Stop it first. localhost: secondarynamenode running as process 8222. Stop it first. jobtracker running as process 8294. Stop it first. localhost: tasktracker running as process 8472. Stop it first.