Three ways to start or stop 09_hadoop and launch scripts

1. Hadoop starts or stops

1 ) The first way

Start separately HDFS and the MapReduce , the command is as follows:

Start:

$ start-dfs.sh

$ start-mapred.sh

Stop it:

$ stop-mapred.sh

$ start-dfs.sh

2 ) Second Way

Start all or stop all

Start: start-all.sh

Boot order: NameNode , DataNode , Secondarynamenode , Jobtracker , Tasktracker

Stop: stop-all.sh

Stop Order: Jobtracker , Tasktracker , NameNode , DataNode , Secondarynamenode

3 ) The Third way of starting

Each daemon is started one by one and the boot sequence is as follows:

NameNode , DataNode , Secondarynamenode , Jobtracker , Tasktracker

The command is as follows:

Start:

hadoop-daemon.sh Start Namenode

hadoop-daemon.sh Start Datanode

hadoop-daemon.sh Start Secondarynamenode

hadoop-daemon.sh Start Jobtracker

hadoop-daemon.sh Start Tasktracker

Stop it:

hadoop-daemon.sh Stop Jobtracker

hadoop-daemon.sh Stop Tasktracker

hadoop-daemon.sh Stop Namenode

hadoop-daemon.sh Stop Datanode

hadoop-daemon.sh Stop Secondarynamenode

2 , analysis start Shell Script

1 ) View start-all.sh Script:

1th:, this Shell script, executed only on the primary node.

2nd: Start First DFS file System daemon, start again MapReduce The daemon of the framework

3rd: Start HDFS when the file system daemon is called, the Start-dfs.shshell To start a script; MapReduce Daemon, call the Start-mapred.shshell script.

2 ) View start-dfs.sh Script:

1th: This script runs on DFS on the primary node of the file system.

2nd: If you start first DataNode daemon, no boot in progress NameNode Daemon before the process, DataNode log file is always connected NameNode error message.

3rd: Start HDFS Order of Daemons

4th: NameNode start, call is hadoop-daemon.sh scripts;

DataNode and the Secondarynamenode the start call is hadoop-daemons.sh script.

Three ways to start or stop 09_hadoop and launch scripts