Briefly understand namenode and datanode

One 、Hadoop Start of

Before reading this article , You should read ：Hadoop Installation and environment construction .

1. First, in the npdev1 Running on the machine hadoop-daemon.sh start namenode, start-up namenode.

2. And then in npfdev2,npfdev3,npfdev4 Run respectively hadoop-daemon.sh start datanode

3. Once the boot is complete , our Hadoop The physical and logical architecture of is as follows ：

npfdev1 Running on the machine namenode, npfdev2,npfdev3,npfdev4 Run respectively datanode.

4. HDFS The structure is as follows ：

Hadoop Simply put, it is backed up by multiple machines , A system that can scale horizontally .

Two 、 Use the command line to view the entire colony Information

1. Use hdfs dfsadmin -report command

We want to better show , Use hdfs dfsadmin -report | more command ：

1.1 see colony Usage situation ：

1.2 see colony Of datanode The situation of , All in all 3 individual live Of datanode.

first datanode:

the second datanode:

Third datanode:

3、 ... and 、 Use web View the whole cluster information on the page

1. visit http://npfdev1:50070, choice Overview Options :

2.  choice Datanodes Options :

 3.  choice Utilities Under the options browse directory, You can see the usage of the whole hard disk :

4.  choice Utilities Under the options logs, You can see the startup log :

Four 、 close Hadoop colony

1. First, in the npdev1 Running on the machine hadoop-daemon.sh stop namenode.

2. And then in npfdev2,npfdev3,npfdev4 Run respectively hadoop-daemon.sh stop datanode.

5、 ... and 、 Startup and shutdown Hadoop Summary of cluster command steps ：

1. stay master Start the hadoop-daemon.sh start namenode. 2. stay slave Start the hadoop-daemon.sh start datanode. 3. use jps Observe the execution result of the instruction . 4. use hdfs dfsadmin -report Observe the cluster configuration . 5. adopt http://npfdev1:50070 Observe the cluster operation on the interface . 6. use hadoop-daemon.sh stop ... Manually shut down the cluster .