Hadoop Ha-12927965

birk

　　一 、Hadoop Ha 安装准备工作1.zookeeper集群masterslave1slave2Hadoop集群master Namenode1 ResourceManager1 Journalnode1slave1 Namenode1 ResourceManager2 Journalnode2slave2 DataNode1slave3 DataNode22.修改主机名hostnamectl set-hostname master（slave1，slave2，slave3）3.修改/etc/hosts文件192.168.197.139  master（slave1，slave2，slave3）4.设置ssh免密码登录ssh-keygen  一直回车ssh-copy-id master（slave1，slave2，slave3）5.安装jdk用xftp上传jdk、zookeeper、Hadoop包到/usr/local下tar xzvf jdk*mv jdk* java 修改环境变量/etc/profile export JAVA_HOME=/usr/local/javaexport JRE_HOME=/usr/java/jreexport>fs.defaultFShdfs://ns1hadoop.tmp.dir/usr/local/hadoop/tmpha.zookeeper.quorummaster:2181,slave1:2181,slave2:2181　　(3)修改hdfs-site.xml
dfs.nameservicesns1dfs.ha.namenodes.ns1nn1,nn1dfs.namenode.rpc-address.ns1.nn1master:9000dfs.namenode.http-address.ns1.nn1master:50070dfs.namenode.rpc-address.ns1.nn2slave1:9000dfs.namenode.http-address.ns1.nn2slave1:50070dfs.namenode.shared.edits.dirqjournal://master:8485;slave1:8485;/ns1dfs.journalnode.edits.dir/usr/local/hadoop/journaldfs.ha.automatic-failover.enabledtruedfs.client.failover.proxy.provider.ns1org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProviderdfs.ha.fencing.methodssshfenceshell(/bin/true)dfs.ha.fencing.ssh.private-key-files/root/.ssh/id_rsadfs.ha.fencing.ssh.connect-timeout30000　　(4)修改mapred-site.xml
mapreduce.framework.nameyarn　　(6)修改yarn-site.xml
yarn.resourcemanager.ha.enabledtrueyarn.resourcemanager.cluster-idyrcyarn.resourcemanager.ha.rm-idsrm1,rm2yarn.resourcemanager.hostname.rm1masteryarn.resourcemanager.hostname.rm2slave1yarn.resourcemanager.zk-addressmaster:2181,slave1:2181,slave2:2181yarn.nodemanager.aux-servicesmapreduce_shuffle　　(6)修改slavesslave2slave34、将配置好的hadoop拷贝到其他节点scp -r /usr/local/hadoop slave1(slave2，slave3）:/usr/local5.scp -r /etc/profile slave1(slave2，slave3）:/etc/profilesource /etc/profile四、启动zookeeper集群zkServer.sh start 打开zookeeperzkServer.sh statuszkServer.sh stop  关闭zookeeperzkServer.sh restart  重启zookeeper五、在master和slave1上启动Journalnodehadoop-daemon.sh start journalnode六、格式化HDFS（master）1. hdfs namenode -format2. 格式化zookeeperhdfs zkfc -formatZK七、在master上启动Hadoop集群start-all.sh