- Hadoop Beginner's Guide
- Garry Turkington
- 324字
- 2021-07-29 16:51:35
Time for action – starting Hadoop
Unlike the local mode of Hadoop, where all the components run only for the lifetime of the submitted job, with the pseudo-distributed or fully distributed mode of Hadoop, the cluster components exist as long-running processes. Before we use HDFS or MapReduce, we need to start up the needed components. Type the following commands; the output should look as shown next, where the commands are included on the lines prefixed by $:
- Type in the first command:
$ start-dfs.sh starting namenode, logging to /home/hadoop/hadoop/bin/../logs/hadoop-hadoop-namenode-vm193.out localhost: starting datanode, logging to /home/hadoop/hadoop/bin/../logs/hadoop-hadoop-datanode-vm193.out localhost: starting secondarynamenode, logging to /home/hadoop/hadoop/bin/../logs/hadoop-hadoop-secondarynamenode-vm193.out
- Type in the second command:
$ jps 9550 DataNode 9687 Jps 9638 SecondaryNameNode 9471 NameNode
- Type in the third command:
$ hadoop dfs -ls / Found 2 items drwxr-xr-x - hadoop supergroup 0 2012-10-26 23:03 /tmp drwxr-xr-x - hadoop supergroup 0 2012-10-26 23:06 /user
- Type in the fourth command:
$ start-mapred.sh starting jobtracker, logging to /home/hadoop/hadoop/bin/../logs/hadoop-hadoop-jobtracker-vm193.out localhost: starting tasktracker, logging to /home/hadoop/hadoop/bin/../logs/hadoop-hadoop-tasktracker-vm193.out
- Type in the fifth command:
$ jps 9550 DataNode 9877 TaskTracker 9638 SecondaryNameNode 9471 NameNode 9798 JobTracker 9913 Jps
What just happened?
The start-dfs.sh command, as the name suggests, starts the components necessary for HDFS. This is the NameNode to manage the filesystem and a single DataNode to hold data. The SecondaryNameNode is an availability aid that we'll discuss in a later chapter.
After starting these components, we use the JDK's jps utility to see which Java processes are running, and, as the output looks good, we then use Hadoop's dfs utility to list the root of the HDFS filesystem.
After this, we use start-mapred.sh to start the MapReduce components—this time the JobTracker and a single TaskTracker—and then use jps again to verify the result.
There is also a combined start-all.sh file that we'll use at a later stage, but in the early days it's useful to do a two-stage start up to more easily verify the cluster configuration.