hadoop部署服务器
系统 | 主机名 | IP |
centos6.9 | hadoop01 | 192.168.72.21 |
centos6.9 | hadoop02 | 192.168.72.22 |
centos6.9 | hadoop03 | 192.168.72.23 |
基础环境准备
1.修改Linux主机名
2.修改IP
3.修改主机名和IP的映射关系 /etc/hosts
4.关闭防火墙
5.ssh免登陆
6.安装JDK,配置环境变量等
7.注意集群时间要同步
8.安装zookeeper集群
部署节点规划
集群部署节点角色的规划(3节点)
server01 namenode resourcemanager zkfc nodemanager datanode zookeeper journal node
server02 namenode resourcemanager zkfc nodemanager datanode zookeeper journal node
server03 datanode nodemanager zookeeper journal node
安装部署
上传编译好的hadoop安装程序到服务器上并解压到指定目录
[root@hadoop01 soft]# tar zxvf spark-2.2.0-bin-2.6.0-cdh5.14.0.tgz -C /usr/local/[root@hadoop01 soft]# cd /usr/local/[root@hadoop01 soft]# mv hadoop-2.6.0-cdh5.14.0 hadoop-HA
配置Hadoop环境变量,编辑/etc/profile添加Hadoop环境变量
[root@hadoop01 soft]# vim /etc/profile
############################HADOOP HA########################################export HADOOP_HOME=/usr/local/hadoop-HAexport PATH=$PATH:$HADOOP_HOME/bin:$HADOOP_HOME/sbin
修改Hadoop相关配置文件
修改hadoop-env.sh文件,配置HAVA_HOME如下
export JAVA_HOME=/usr/local/java/jdk1.8.0_201
修改core-site.xml
fs.defaultFS hdfs://cluster1 hadoop.tmp.dir /usr/local/hadoop-HA/tmp ha.zookeeper.quorum hadoop01:2181,hadoop02:2181,hadoop03:2181
修改hdfs-site.xml
dfs.nameservices cluster1 dfs.ha.namenodes.cluster1 nn1,nn2 dfs.namenode.rpc-address.cluster1.nn1 hadoop01:9000 dfs.namenode.http-address.cluster1.nn1 hadoop1:50070 dfs.namenode.rpc-address.cluster1.nn2 hadoop02:9000 dfs.namenode.http-address.cluster1.nn2 hadoop02:50070 dfs.namenode.shared.edits.dir qjournal://hadoop01:8485;hadoop02:8485;hadoop03:8485/cluster1 dfs.journalnode.edits.dir /usr/local/hadoop-HA/journaldata dfs.ha.automatic-failover.enabled true dfs.client.failover.proxy.provider.cluster1 org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider dfs.ha.fencing.methods sshfence dfs.ha.fencing.ssh.private-key-files /home/hadoop/.ssh/id_rsa dfs.ha.fencing.ssh.connect-timeout 30000
修改mapred-site.xml
mapreduce.framework.name yarn
修改yarn-site.xml
yarn.resourcemanager.ha.enabled true yarn.resourcemanager.cluster-id yrc yarn.resourcemanager.ha.rm-ids rm1,rm2 yarn.resourcemanager.hostname.rm1 hadoop01 yarn.resourcemanager.hostname.rm2 hadoop02 yarn.resourcemanager.zk-address hadoop01:2181,hadoop02:2181,hadoop03:2181 yarn.nodemanager.aux-services mapreduce_shuffle
修改slaves(slaves是指定子节点的位置,因为要在hadoop01上启动HDFS、在hadoop03启动yarn,所以hadoop01上的slaves文件指定的是datanode的位置,hadoop03上的slaves文件指定的是nodemanager的位置)
hadoop01hadoop02hadoop03
Hadoop集群启动
启动zookeeper集群(分别在hadoop01、hadoop02、hadoop03上启动zk)
bin/zkServer.sh start
#查看状态:一个leader,两个follower
bin/zkServer.sh status
手动启动journalnode(分别在hadoop01、hadoop02、hadoop03上执行)
hadoop-daemon.sh start journalnode #运行jps命令检验,hadoop01、hadoop02、hadoop03上多了JournalNode进程
格式化namenode
在hadoop00上执行命令:
hdfs namenode -format #格式化后会在根据core-site.xml中的hadoop.tmp.dir配置的目录下生成个hdfs初始化文件
把hadoop.tmp.dir配置的目录下所有文件拷贝到另一台namenode节点所在的机器
scp -r hadoop-HA root@hadoop02:$PWD
##也可以这样,建议hdfs namenode -bootstrapStandby
格式化ZKFC(在active上执行即可)
hdfs zkfc -formatZK
启动HDFS(在hadoop00上执行)
start-dfs.sh
启动YARN
start-yarn.sh
还需要手动在standby上手动启动备份的 resourcemanager
yarn-daemon.sh start resourcemanager
JPS查看启动进程
集群验证
到此,hadoop-2.6.4配置完毕,可以统计浏览器访问:
NameNode 'hadoop01:9000' (active)
NameNode 'hadoop02:9000' (standby)
验证HDFS HA
首先向hdfs上传一个文件
hadoop fs -put /etc/profile /profile
hadoop fs -ls /
然后再kill掉active的NameNode
kill -9 <pid of NN>
通过浏览器访问:http://192.168.1.202:50070
NameNode 'hadoop02:9000' (active)
这个时候hadoop02上的NameNode变成了active
在执行命令:
hadoop fs -ls /
-rw-r--r-- 3 root supergroup 1926 2014-02-06 15:36 /profile
刚才上传的文件依然存在!!!
手动启动那个挂掉的NameNode
hadoop-daemon.sh start namenode
通过浏览器访问:http://192.168.1.201:50070
NameNode 'hadoop01:9000' (standby)
验证YARN:
运行一下hadoop提供的demo中的WordCount程序:
hadoop jar share/hadoop/mapreduce/hadoop-mapreduce-examples-2.4.1.jar wordcount /profile /out
测试集群工作状态的一些指令 :
hdfs dfsadmin -report 查看hdfs的各节点状态信息
cluster1n/hdfs haadmin -getServiceState nn1 获取一个namenode节点的HA状态
scluster1n/hadoop-daemon.sh start namenode 单独启动一个namenode进程
./hadoop-daemon.sh start zkfc 单独启动一个zkfc进程