three ways to start and close Hadoop's five daemon processes
The first type of startup:
Go to "hadoop-1.x/bin" directory, perform START-ALL.SH,JPS view process, start all successfully.
19043 NameNode
19156 DataNode
19271SecondaryNameNode
19479TaskTracker
24008 Jps
19353JobTracker
View start-all.sh Code Discovery:
# Start Dfs daemons
"$bin"/start-dfs.sh--config$hadoop_conf_dir
# Start Mapred daemons
"$bin"/start-mapred.sh--config$hadoop_conf_dir
start-all.sh boot is dependent on, start-dfs.sh and start-mapred.sh two boot files, it is necessary to note that the order of execution, is executed first, start-dfs.sh then execute start-mapred.sh
The second mode of startup:
Go to "hadoop-1.x/bi" directory, execute start-dfs.sh and then execute START-MAPRED.SH,JPS view process, all start successfully.
19043 NameNode
19156 DataNode
19271SecondaryNameNode
19479TaskTracker
24008 Jps
19353JobTracker
See START-DFS.SH and start-mapred.sh Code discovery:
"Start-dfs.sh"
# Start Dfsdaemons
# Startnamenode after datanodes, to minimize time Namenode was up w/o data
# note:datanodes would log connection errors until Namenode starts
"$bin"/hadoop-daemon.sh--config $HADOOP _conf_dir start Namenode $nameStartOpt
"$bin"/hadoop-daemons.sh--config $HADOOP _conf_dir start Datanode $dataStartOpt
"$bin"/hadoop-daemons.sh--config $HADOOP _conf_dir--hosts Masters Start Secondarynamenode
"Start-mapred.sh"
# Start Mapreddaemons
# Startjobtracker first to minimize connection errors at startup
"$bin"/hadoop-daemon.sh--config $HADOOP _conf_dir start Jobtracker
"$bin"/hadoop-daemons.sh--config $HADOOP _conf_dir start Tasktracker
It is noteworthy that the start-dfs.sh and start-mapred.sh start-up is dependent on hadoop-daemon.sh and hadoop-daemons.sh files to start;
Note the order in which the boot order is Namenode, Datanode, Secondarynamenode, Jobtracker, Tasktracker
The Third Way to start:
Enter the "Hadoop-1.x/bi" directory with the following command:
"$./hadoop-daemon.sh Namenode"
"$./hadoop-daemon.sh Datanode"
"$./hadoop-daemon.sh Secondarynamenode"
"$./hadoop-daemon.sh Jobtracker"
"$./hadoop-daemon.sh Tasktracker"
JPS view process, start all successfully
19043 NameNode
19156 DataNode
19271SecondaryNameNode
19479TaskTracker
24008 Jps
19353JobTracker
The first method of closing:
Call stop-all.sh, same principle
The second mode of management:
Call stop-mapred.sh and call stop-dfs.sh to pay attention to the order of execution, the same principle
The Third Way of closing:
The closing order is: Jobtracker, Tasktracker, Namenode, Datanode, Secondarynamenode
Enter the "Hadoop-1.x/bi" directory with the following command:
"$./hadoop-daemon.sh Jobtracker"
"$./hadoop-daemon.sh Tasktracker"
"$./hadoop-daemon.sh Namenode"
"$./hadoop-daemon.sh Datanode"
"$./hadoop-daemon.sh Secondarynamenode"
It should be noted that the order of execution must be strictly observed.
These are all the three ways to start Hadoop's five daemon processes. The individual sections refer to the Cloud Sail Big Data Video course. If there are errors or problems, comment comments are actively discussed together.