hadoop-spark集群安装
一、准备1.上传hadoop-2.7.3.tar.gz到node01:/tools
2.tar -zxvf /ren/hadoop-2.7.3.tar.gz -C /ren/
二、配置
1.修改变量
vi etc/profile
export HADOOP_HOME=/ren/hadoop-2.7.3
export PATH=$PATH:$HADOOP_HOME/bin:$HADOOP_HOME/sbin
source /etc/profile
cd /ren/hadoop-2.7.3/etc/hadoop开始修改配置
2.修改 hadoop-env.sh
vi hadoop-env.sh
export JAVA_HOME=/usr/java/jdk1.8.0_101
3.修改salves
vi salve
添加 node01
node02
node03
4.修改core-site.xml
<configuration>
<property>
<!--///hdfs的nameservice -->
<name>fs.defaultFS</name>
<value>hdfs://ren</value>
</property>
<property>
<!--//zookeeper地址 -->
<name>ha.zookeeper.quorum</name>
<value>node01:2181,node02:2181,node03:2181</value>
</property>
<property>
<!--//hadoop临时文件目录 -->
<name>hadoop.tmp.dir</name>
<value>/ren/hadoop-2.7.3/tmpdir</value>
</property>
</configuration>
5.修改hdfs-site.xml
<configuration>
<!-- dfs.nameservices 命名空间的逻辑名称，多个用,分割 -->
<property>
<name>dfs.nameservices</name>
<value>ren</value>
</property>
<!-- 指定ren下有两个namenode，分别是nn1,nn2 -->
<property>
<name>dfs.ha.namenodes.ren</name>
<value>nn1,nn2</value>
</property>
<!-- 指定nn1的RPC通信地址 -->
<property>
<name>dfs.namenode.rpc-address.ren.nn1</name>
<value>node01:8020</value>
</property>
<!-- 指定nn1的HTTP通信地址 -->
<property>
<name>dfs.namenode.http-address.ren.nn1</name>
<value>node01:50070</value>
</property>
<!-- 指定nn2的RPC通信地址 -->
<property>
<name>dfs.namenode.rpc-address.ren.nn2</name>
<value>node02:8020</value>
</property>
<!-- 指定nn2的HTTP通信地址 -->
<property>
<name>dfs.namenode.http-address.ren.nn2</name>
<value>node02:50070</value>
</property>
<!-- 指定namenode的元数据存放的Journal Node的地址，必须基数，至少三个 -->
<property>
<name>dfs.namenode.shared.edits.dir</name>
<value>qjournal://node01:8485;node02:8485;node03:8485/ren</value>
</property>
<!-- journalnode地址-->
<property>
<name>dfs.journalnode.edits.dir</name>
<value>/ren/hadoop-2.7.3/journal</value>
</property>
<!-- namenode启动失败，允许自动切换-->
<property>
<name>dfs.ha.automatic-failover.enabled</name>
<value>true</value>
</property>
<!-- 自动切换方式-->
<property>
<name>dfs.client.failover.proxy.provider.ren</name>
<value>org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider</value>
</property>
<!-- 隔离机制方法-->
<property>
<name>dfs.ha.fencing.methods</name>
<value>sshfence</value>
</property>
<!-- 配置隔离需要ssh免密码-->
<property>
<name>dfs.ha.fencing.ssh.private-key-files</name>
<value>/root/.ssh/id_rsa</value>
</property>
</configuration>
6.修改mapred-site.xml
mv mapred-site.xml.template mapred-site.xml
vi mapred-site.xml
修改
<configuration>
<!-- mr框架由yarn管理-->
<property>
<name>mapreduce.framework.name</name>
<value>yarn</value>
</property>
</configuration>
7.修改yarn-site.xml
<configuration>
<!-- 允许高可用-->
<property>
<name>yarn.resourcemanager.ha.enabled</name>
<value>true</value>
</property>
<!--指定mr集群id-->
<property>
<name>yarn.resourcemanager.cluster-id</name>
<value>ren_mr</value>
</property>
<!-- 指定多个resourcemanager的id-->
<property>
<name>yarn.resourcemanager.ha.rm-ids</name>
<value>rm1,rm2</value>
</property>
<!-- 分别指定rm的hostname-->
<property>
<name>yarn.resourcemanager.hostname.rm1</name>
<value>node01</value>
</property>
<property>
<name>yarn.resourcemanager.hostname.rm2</name>
<value>node02</value>
</property>
<!-- 指定zk地址-->
<property>
<name>yarn.resourcemanager.zk-address</name>
<value>node01:2181,node02:2181,node03:2181</value>
</property>
<!-- 指定reducer取数据方式为mapreduce_shuffle-->
<property>
<name>yarn.nodemanager.aux-services</name>
<value>mapreduce_shuffle</value>
</property>
</configuration>
三、启动
1.同步
scp -r /ren/hadoop-2.7.3 root@node02:/ren/
scp -r /ren/hadoop-2.7.3 root@node03:/ren/
2. 在各个journalnode节点启动journalnode
hadoop-daemon.sh start journalnode
3.选择一个NN节点，格式化，启动
hdfs namenode -format
hadoop-daemon.sh start namenode
4.把格式化后的meta data同步到另一个NN，启动
hdfs namenode -bootstrapStandby
hadoop-daemon.sh start namenode
5.在任意一个NN节点初始化ZK
hdfs zkfc -formatZK
6.启动集群
start-all.sh
关闭集群：stop-all.sh
查看namenode状态：http://noed01:50070
查看yarn状态：http://node01:8088
页:
[1]