上接:hadoop伪分布式安装
注意:要关闭操作系统的防火墙
概述
本文根据官网实现:
https://hadoop.apache.org/docs/r3.0.0/hadoop-project-dist/hadoop-common/SingleCluster.html
修改 mapred-site.xml
vim etc/hadoop/mapred-site.xml
指定MapReduce程序运行在Yarn上,配置如下:
<property>
<name>mapreduce.framework.name</name>
<value>yarn</value>
</property>
指定mr运行在yarn上
保存
修改yarn-site.xml
vim etc/hadoop/yarn-site.xml
指定获取数据的方式
<property>
<name>yarn.nodemanager.aux-services</name>
<value>mapreduce_shuffle</value>
</property>
不检查虚拟内存
关闭检查内存配置,原因见 链接,操作如下:
<property>
<name>yarn.nodemanager.vmem-check-enabled</name>
<value>false</value>
</property>
设置classpath
在hadoop3.x中必须设置,否则执行 hadoop-mapreduce-examples-3.0.3.jar 例子报错
执行下面命令,查看classpath:
hadoop classpath
执行结果如下:
/program/hadoop-3.0.3/etc/hadoop:/program/hadoop-3.0.3/share/hadoop/common/lib/*:/program/hadoop-3.0.3/share/hadoop/common/*:/program/hadoop-3.0.3/share/hadoop/hdfs:/program/hadoop-3.0.3/share/hadoop/hdfs/lib/*:/program/hadoop-3.0.3/share/hadoop/hdfs/*:/program/hadoop-3.0.3/share/hadoop/mapreduce/lib/*:/program/hadoop-3.0.3/share/hadoop/mapreduce/*:/program/hadoop-3.0.3/share/hadoop/yarn:/program/hadoop-3.0.3/share/hadoop/yarn/lib/*:/program/hadoop-3.0.3/share/hadoop/yarn/*
在 yarn-site.xml 添加如下:
<property>
<name>yarn.application.classpath</name>
<value>/program/hadoop-3.0.3/etc/hadoop:/program/hadoop-3.0.3/share/hadoop/common/lib/*:/program/hadoop-3.0.3/share/hadoop/common/*:/program/hadoop-3.0.3/share/hadoop/hdfs:/program/hadoop-3.0.3/share/hadoop/hdfs/lib/*:/program/hadoop-3.0.3/share/hadoop/hdfs/*:/program/hadoop-3.0.3/share/hadoop/mapreduce/lib/*:/program/hadoop-3.0.3/share/hadoop/mapreduce/*:/program/hadoop-3.0.3/share/hadoop/yarn:/program/hadoop-3.0.3/share/hadoop/yarn/lib/*:/program/hadoop-3.0.3/share/hadoop/yarn/*</value>
</property>
value 部分是 hadoop classpath 执行的结果
保存
启动
start-yarn.sh
web访问
注意:要关闭操作系统的防火墙,否则无法访问
停止
stop-yarn.sh
