Hadoop03 Hdfs完整集群搭建

部署分布

 NameNode 是基于内存 最好单独部署

node1 NameNode DataNode

node2 DataNode

node3 DataNode

node4 SNN

node 2 node3 node4 分别部署Jdk 和环境变量HADOOP_HOME  必须跟node1一样

修改配置文件

 node1修改 文件

vim hdfs-site.xml

<configuration>
        <!-- 副本数量为2-->
        <property>
        <name>dfs.replication</name>
        <value>2</value>
        </property>
<!-- 修改默认NameNode持久化目录 防止内存紧缺时回收/tmp下文件  -->
        <property>
                <name>dfs.namenode.name.dir</name>
                <value>/data/DataOrLogs/hadoop/full/hadoop/local/dfs/name</value>
        </property>

        <!-- 修改默认DataNode持久化目录 防止内存紧缺时回收/tmp下文件  -->
        <property>
                <name>dfs.datanode.data.dir</name>
                <value>/data/DataOrLogs/hadoop/full/hadoop/local/dfs/data</value>
        </property>

 <!-- SecondaryNameNode 放在Node4 -->
         <property>
                <name>dfs.namenode.secondary.http-address</name>
                <value>node4:50090</value>
        </property>
                <property>
                <name>dfs.namenode.checkpoint.dir</name>
                <value>/data/DataOrLogs/hadoop/full/hadoop/local/dfs/secondary</value>
        </property>
</configuration>

vim slaves

node1
node2
node3

复制到node2 node3 node4 

scp -r hadoop-2.6.5/ root@node2:/usr/local/soft/hadoop-2.6.5/
scp -r hadoop-2.6.5/ root@node3:/usr/local/soft/hadoop-2.6.5/
scp -r hadoop-2.6.5/ root@node4:/usr/local/soft/hadoop-2.6.5/

启动

格式化

hdfs namenode -format 

启动 

start-dfs.sh