Hadoop用户重新部署HDFS

Posted 2020-11-06

前言：
在这篇文章中https://www.jianshu.com/p/eeae2f37a48c
我们使用的是root用户来部署的，在生产环境中，一般某个组件是由某个用户来启动的，本篇文章介绍下怎样用hadoop用户来重新部署伪分布式(HDFS)

1.前期准备

创建hadoop用户，及配置ssh免密登录
参考：https://www.jianshu.com/p/589bb43e0282

2.停止root启动的HDFS进程并删除/tmp目录下的存储文件

[[email protected] hadoop-2.8.1]# pwd
/opt/software/hadoop-2.8.1
[[email protected] hadoop-2.8.1]# jps
32244 NameNode
32350 DataNode
32558 SecondaryNameNode
1791 Jps
[[email protected] hadoop-2.8.1]# sbin/stop-dfs.sh 
Stopping namenodes on [hadoop000]
hadoop000: stopping namenode
localhost: stopping datanode
Stopping secondary namenodes [0.0.0.0]
0.0.0.0: stopping secondarynamenode
[[email protected] hadoop-2.8.1]# jps
2288 Jps
[[email protected]  hadoop-2.8.1]# rm -rf /tmp/hadoop-* /tmp/hsperfdata_*

3.更改文件属主

[[email protected] software]# pwd
/opt/software
[[email protected] software]# chown -R hadoop:hadoop hadoop-2.8.1

4.进入hadoop用户 修改相关配置文件

#第一步:
[[email protected] hadoop]$ pwd
/opt/software/hadoop-2.8.1/etc/hadoop
[[email protected] hadoop]$ vi hdfs-site.xml 
<configuration&gt;
<property&gt;
<name&gt;dfs.replication&lt;/name&gt;
<value&gt;1&lt;/value&gt;
</property&gt;
<property&gt;
<name&gt;dfs.namenode.secondary.http-address&lt;/name&gt;
<value&gt;192.168.6.217:50090&lt;/value&gt;
</property&gt;
<property&gt;
<name&gt;dfs.namenode.secondary.https-address&lt;/name&gt;
<value&gt;192.168.6.217:50091&lt;/value&gt;
</property&gt;
</configuration&gt;
#第二步：
[[email protected] hadoop]$ vi core-site.xml 
<configuration&gt;
<property&gt;
<name&gt;fs.defaultFS&lt;/name&gt;
<value&gt;hdfs://192.168.6.217:9000&lt;/value&gt;
</property&gt;
</configuration&gt;
#第三步：
[[email protected] hadoop]# vi slaves 
192.168.6.217

5.格式化和启动

[[email protected] hadoop-2.8.1]$ pwd
/opt/software/hadoop-2.8.1
[[email protected] hadoop-2.8.1]$ bin/hdfs namenode -format
[[email protected] hadoop-2.8.1]$ sbin/start-dfs.sh
Starting namenodes on [hadoop000]
hadoop000: starting namenode, logging to /opt/software/hadoop-2.8.1/logs/hadoop-hadoop-namenode-hadoop000.out
192.168.6.217: starting datanode, logging to /opt/software/hadoop-2.8.1/logs/hadoop-hadoop-datanode-hadoop000.out
Starting secondary namenodes [hadoop000]
hadoop000: starting secondarynamenode, logging to /opt/software/hadoop-2.8.1/logs/hadoop-hadoop-secondarynamenode-hadoop000.out
[[email protected] hadoop-2.8.1]$ jps
3141 Jps
2806 DataNode
2665 NameNode
2990 SecondaryNameNode
#至此发现HDFS三个进程都是以hadoop000启动，