假如我们只有3台linux虚拟机，主机名分别为hadoop01、hadoop02和hadoop03，在这3台机器上，hadoop集群的部署情况如下：
hadoop01：1个namenode，1个datanode，1个journalnode，1个zkfc，1个resourcemanager，1个nodemanager； hadoop02：1个namenode，1个datanode，1个journalnode，1个zkfc，1个resourcemanager，1个nodemanager； hadoop03：1个datenode，1个journalnode，1个nodemanager；
下面我们来介绍启动hdfs和yarn的一些命令。
1.启动hdfs集群（使用hadoop的批量启动脚本）
/root/apps/hadoop/sbin/start-dfs.sh
[root@hadoop01 ~]# /root/apps/hadoop/sbin/start-dfs.sh Starting namenodes on [hadoop01 hadoop02] hadoop01: starting namenode, logging to /root/apps/hadoop/logs/hadoop-root-namenode-hadoop01.out hadoop02: starting namenode, logging to /root/apps/hadoop/logs/hadoop-root-namenode-hadoop02.out hadoop03: starting datanode, logging to /root/apps/hadoop/logs/hadoop-root-datanode-hadoop03.out hadoop02: starting datanode, logging to /root/apps/hadoop/logs/hadoop-root-datanode-hadoop02.out hadoop01: starting datanode, logging to /root/apps/hadoop/logs/hadoop-root-datanode-hadoop01.out Starting journal nodes [hadoop01 hadoop02 hadoop03] hadoop03: starting journalnode, logging to /root/apps/hadoop/logs/hadoop-root-journalnode-hadoop03.out hadoop02: starting journalnode, logging to /root/apps/hadoop/logs/hadoop-root-journalnode-hadoop02.out hadoop01: starting journalnode, logging to /root/apps/hadoop/logs/hadoop-root-journalnode-hadoop01.out Starting ZK Failover Controllers on NN hosts [hadoop01 hadoop02] hadoop01: starting zkfc, logging to /root/apps/hadoop/logs/hadoop-root-zkfc-hadoop01.out hadoop02: starting zkfc, logging to /root/apps/hadoop/logs/hadoop-root-zkfc-hadoop02.out [root@hadoop01 ~]#
从上面的启动日志可以看出，start-dfs.sh这个启动脚本是通过ssh对多个节点的namenode、datanode、journalnode以及zkfc进程进行批量启动的。
2.停止hdfs集群（使用hadoop的批量启动脚本）
/root/apps/hadoop/sbin/stop-dfs.sh
[root@hadoop01 ~]# /root/apps/hadoop/sbin/stop-dfs.sh Stopping namenodes on [hadoop01 hadoop02] hadoop02: stopping namenode hadoop01: stopping namenode hadoop02: stopping datanode hadoop03: stopping datanode hadoop01: stopping datanode Stopping journal nodes [hadoop01 hadoop02 hadoop03] hadoop03: stopping journalnode hadoop02: stopping journalnode hadoop01: stopping journalnode Stopping ZK Failover Controllers on NN hosts [hadoop01 hadoop02] hadoop01: stopping zkfc hadoop02: stopping zkfc [root@hadoop01 ~]#
3.启动单个进程
[root@hadoop01 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh start namenode starting namenode, logging to /root/apps/hadoop/logs/hadoop-root-namenode-hadoop01.out
[root@hadoop02 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh start namenode starting namenode, logging to /root/apps/hadoop/logs/hadoop-root-namenode-hadoop02.out
[root@hadoop01 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh start datanode starting datanode, logging to /root/apps/hadoop/logs/hadoop-root-datanode-hadoop01.out
[root@hadoop02 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh start datanode starting datanode, logging to /root/apps/hadoop/logs/hadoop-root-datanode-hadoop02.out
[root@hadoop03 apps]# /root/apps/hadoop/sbin/hadoop-daemon.sh start datanode starting datanode, logging to /root/apps/hadoop/logs/hadoop-root-datanode-hadoop03.out
[root@hadoop01 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh start journalnode starting journalnode, logging to /root/apps/hadoop/logs/hadoop-root-journalnode-hadoop01.out
[root@hadoop02 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh start journalnode starting journalnode, logging to /root/apps/hadoop/logs/hadoop-root-journalnode-hadoop02.out
[root@hadoop03 apps]# /root/apps/hadoop/sbin/hadoop-daemon.sh start journalnode starting journalnode, logging to /root/apps/hadoop/logs/hadoop-root-journalnode-hadoop03.out
[root@hadoop01 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh start zkfc starting zkfc, logging to /root/apps/hadoop/logs/hadoop-root-zkfc-hadoop01.out
[root@hadoop02 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh start zkfc starting zkfc, logging to /root/apps/hadoop/logs/hadoop-root-zkfc-hadoop02.out
分别查看启动后3台虚拟机上的进程情况：
[root@hadoop01 ~]# jps 6695 DataNode 2002 QuorumPeerMain 6879 DFSZKFailoverController 7035 Jps 6800 JournalNode 6580 NameNode [root@hadoop01 ~]#
[root@hadoop02 ~]# jps 6360 JournalNode 6436 DFSZKFailoverController 2130 QuorumPeerMain 6541 Jps 6255 DataNode 6155 NameNode [root@hadoop02 ~]#
[root@hadoop03 apps]# jps 5331 Jps 5103 DataNode 5204 JournalNode 2258 QuorumPeerMain [root@hadoop03 apps]#
3.停止单个进程
[root@hadoop01 ~]# jps 6695 DataNode 2002 QuorumPeerMain 8486 Jps 6879 DFSZKFailoverController 6800 JournalNode 6580 NameNode [root@hadoop01 ~]# [root@hadoop01 ~]# [root@hadoop01 ~]# [root@hadoop01 ~]# [root@hadoop01 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh stop zkfc stopping zkfc [root@hadoop01 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh stop journalnode stopping journalnode [root@hadoop01 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh stop datanode stopping datanode [root@hadoop01 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh stop namenode stopping namenode [root@hadoop01 ~]# jps 2002 QuorumPeerMain 8572 Jps [root@hadoop01 ~]#
[root@hadoop02 ~]# jps 6360 JournalNode 6436 DFSZKFailoverController 2130 QuorumPeerMain 7378 Jps 6255 DataNode 6155 NameNode [root@hadoop02 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh stop zkfc stopping zkfc [root@hadoop02 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh stop journalnode stopping journalnode [root@hadoop02 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh stop datanode stopping datanode [root@hadoop02 ~]# /root/apps/hadoop/sbin/hadoop-daemon.sh stop namenode stopping namenode [root@hadoop02 ~]# jps 7455 Jps 2130 QuorumPeerMain [root@hadoop02 ~]#
[root@hadoop03 apps]# jps 5103 DataNode 5204 JournalNode 5774 Jps 2258 QuorumPeerMain [root@hadoop03 apps]# /root/apps/hadoop/sbin/hadoop-daemon.sh stop journalnode stopping journalnode [root@hadoop03 apps]# /root/apps/hadoop/sbin/hadoop-daemon.sh stop datanode stopping datanode [root@hadoop03 apps]# jps 5818 Jps 2258 QuorumPeerMain [root@hadoop03 apps]#
3.启动yarn集群（使用hadoop的批量启动脚本）
/root/apps/hadoop/sbin/start-yarn.sh
[root@hadoop01 ~]# /root/apps/hadoop/sbin/start-yarn.sh starting yarn daemons starting resourcemanager, logging to /root/apps/hadoop/logs/yarn-root-resourcemanager-hadoop01.out hadoop03: starting nodemanager, logging to /root/apps/hadoop/logs/yarn-root-nodemanager-hadoop03.out hadoop02: starting nodemanager, logging to /root/apps/hadoop/logs/yarn-root-nodemanager-hadoop02.out hadoop01: starting nodemanager, logging to /root/apps/hadoop/logs/yarn-root-nodemanager-hadoop01.out [root@hadoop01 ~]#
从上面的启动日志可以看出，start-yarn.sh启动脚本只在本地启动一个ResourceManager进程，而3台机器上的nodemanager都是通过ssh的方式启动的。所以hadoop02机器上的ResourceManager需要我们手动去启动。
4.启动hadoop02上的ResourceManager进程
/root/apps/hadoop/sbin/yarn-daemon.sh start resourcemanager
5.停止yarn
/root/apps/hadoop/sbin/stop-yarn.sh
[root@hadoop01 ~]# /root/apps/hadoop/sbin/stop-yarn.sh stopping yarn daemons stopping resourcemanager hadoop01: stopping nodemanager hadoop03: stopping nodemanager hadoop02: stopping nodemanager no proxyserver to stop [root@hadoop01 ~]#
通过上面的停止日志可以看出，stop-yarn.sh脚本只停止了本地的那个ResourceManager进程，所以hadoop02上的那个resourcemanager我们需要单独去停止。
6.停止hadoop02上的resourcemanager
/root/apps/hadoop/sbin/yarn-daemon.sh stop resourcemanager
注意：启动和停止单个hdfs相关的进程使用的是"hadoop-daemon.sh"脚本，而启动和停止yarn使用的是"yarn-daemon.sh"脚本。