hadoop集群之增加三台虚拟机进入集群

来源：互联网发布：铁观音淘宝推荐编辑：程序博客网时间：2024/05/21 06:10

在上回书说道我们已经安装4台主机模拟Hadoop工作环境:

详情请见:http://blog.csdn.net/sinat_27902055/article/details/52998442

通过修改/etc/hosts和hadoop的slaves再文件添加三台虚拟机

主机IP地址安装软件进程账号密码master192.168.40.100hadoop/jdknm SecondaryNameNode
hadoop/hadoopweekend001192.168.40.101hadoop/jdkdnhadoop/hadoopweekend002192.168.40.102hadoop/jdkdnhadoop/hadoopweekend003192.168.40.103hadoop/jdkdnhadoop/hadoopweekend004192.168.40.104hadoop/jdkdn zookeeperhadoop/hadoopweekend005192.168.40.105hadoop/jdkdn zookeeperhadoop/hadoopweekend006192.168.40.106hadoop/jdkdn zookeeperhadoop/hadoop

完成配置后启动hadoop集群(不必再格式化namenode会导致无法启动datanode)

在namenode中进行datanode的刷新

sh hadoop-daemon.sh start datanode

在datanode中进行负载均衡操作

start-balancer.sh

再次检查jps和50070确认服务都起来了

如果出现nodemanager stop it first 没有nodemanager进程

先执行stop-all.sh下，然后再执行start-all.sh

----------------------------------------------------------------------------------------------------------------------------------------------------------------------------

0 0