hadoop +hbase +zookeeper 完全分布搭建 (版本一)

xiaoxiao2026-01-29 7

hadoop2.0已经发布了稳定版本了，增加了很多特性，比如HDFS HA、YARN等。最新的hadoop-2.6.0又增加了YARN HA 注意：apache提供的hadoop-2.6.0的安装包是在32位操作系统编译的，因为hadoop依赖一些C++的本地库，所以如果在64位的操作上安装hadoop-2.6.0就需要重新在64操作系统上重新编译

一.重新编译

原因是hadoop-2.6.0.tar.gz安装包是在32位机器上编译的，64位的机器加载本地库.so文件时出错，不影响使用。解决： 1、重新编译源码后将新的lib/native替换到集群中原来的lib/native 2、修改hadoop-env.sh ，增加 export HADOOP_OPTS="-D java.library.path=$HADOOP_PREFIX/lib:$HADOOP_PREFIX/lib/native" 需要进行编译准备工作： 0.安装JDK,使用 java -version 查看jdk版本，确定JDK版本是64位。 a. 解压jdk $ tar -xvzf jdk-7u60-linux-x64.tar.gz b. 设置环境变量 vim /etc/profile export JAVA_HOME=/usr/local/jdk1.7 export HADOOP_HOME=/usr/local/hadoop-2.6.0 export MAVEN_HOME=/opt/apache-maven export FINDBUG_HOME=/opt/findbugs-3.0.0 export ANT_HOME=/opt/apache-ant-1.9.4 export PATH=$PATH:$JAVA_HOME/bin:$HADOOP_HOME/sbin:$HADOOP_HOME/bin :$MAVEN_HOME/bin:$FINDBUG_HOME/bin:$ANT_HOME/bin (备注：不能换行) c.使配置文件生效 $ source /etc/profile 1.安装gcc|gc++ yum install gcc yum install gcc-c++ 验证 2.安装Apache-Maven。 tar -zxvf apache-maven-3.2.1.tar.gz 配置环境变量 vim /etc/profile export MAVEN_HOME=/opt/apache-maven export PATH=.:PATH:$MAVEN_HOME/bin 验证： mvn --version 3.安装Apache-ant(重要). tar -zxvf apache-ant-1.9.4-bin.tar.gz 配置环境变量 vim /etc/profile export MAVEN_HOME=/opt/apache-maven export PATH=$PATH:$ MAVEN_HOME/bin 3.安装protobuf(goole序列化工具) tar -zxvf protobuf-2.5.0.tar.gz ./configuration make #编译 make install 验证：protoc --version 4.安装CMake2.6 or newer 安装 yum install cmake 安装 yum install openssl-devel 安装 yum install ncurses-devel 验证：cmake --version 5.安装make yum install make 验证： make --version 6.Hadoop - hadoop-common-project中的pom.xml添加依赖（hadoop-2.2.0需要修改，hadoop2.6.0版本不需要） <dependency> <groupId>org.mortbay.jetty</groupId> <artifactId>jetty-util</artifactId> <scope>test</scope> </dependency> 7.在编译之前防止 java.lang.OutOfMemoryError: Java heap space 堆栈问题,在centos系统中执行命令： $ export MAVEN_OPTS="-Xms256m -Xmx512m" 8.解压压缩包 tar -zxvf hadoop-2.6.0-src.tar.gz a.执行命令 cd ${hostname_Local}/hadoop-2.6.0/ 目录下 b.编译 mvn package -DskipTests -Pdist,native c. 编译好的项目放在 hadoop-2.6.0-src/hadoop-dist/target目录下。 /root/Downloads/hadoop-2.6.0-src/hadoop-dist/target 即 hadoop-2.6.0就是编译好的包。 ===================================================================== 编译日志： [INFO] --- maven-jar-plugin:2.3.1:jar (default-jar) @ hadoop-dist --- [WARNING] JAR will be empty - no content was marked for inclusion! [INFO] Building jar: /root/Downloads/hadoop-2.6.0-src/hadoop-dist/target/hadoop-dist-2.6.0.jar [INFO] [INFO] --- maven-source-plugin:2.1.2:jar-no-fork (hadoop-java-sources) @ hadoop-dist --- [INFO] No sources in project. Archive not created. [INFO] [INFO] --- maven-source-plugin:2.1.2:test-jar-no-fork (hadoop-java-sources) @ hadoop-dist --- [INFO] No sources in project. Archive not created. [INFO] [INFO] --- maven-site-plugin:3.3:attach-descriptor (attach-descriptor) @ hadoop-dist --- [INFO] [INFO] --- maven-antrun-plugin:1.7:run (tar) @ hadoop-dist --- [INFO] Executing tasks main: [INFO] Executed tasks [INFO] [INFO] --- maven-javadoc-plugin:2.8.1:jar (module-javadocs) @ hadoop-dist --- [INFO] Building jar: /root/Downloads/hadoop-2.6.0-src/hadoop-dist/target/hadoop-dist-2.6.0-javadoc.jar [INFO] ------------------------------------------------------------------------ [INFO] Reactor Summary: [INFO] [INFO] Apache Hadoop Main ................................. SUCCESS [ 13.582 s] [INFO] Apache Hadoop Project POM .......................... SUCCESS [ 9.846 s] [INFO] Apache Hadoop Annotations .......................... SUCCESS [ 24.408 s] [INFO] Apache Hadoop Assemblies ........................... SUCCESS [ 1.967 s] [INFO] Apache Hadoop Project Dist POM ..................... SUCCESS [ 6.443 s] [INFO] Apache Hadoop Maven Plugins ........................ SUCCESS [ 20.692 s] [INFO] Apache Hadoop MiniKDC .............................. SUCCESS [ 14.250 s] [INFO] Apache Hadoop Auth ................................. SUCCESS [ 23.716 s] [INFO] Apache Hadoop Auth Examples ........................ SUCCESS [ 13.714 s] [INFO] Apache Hadoop Common ............................... SUCCESS [08:46 min] [INFO] Apache Hadoop NFS .................................. SUCCESS [ 47.127 s] [INFO] Apache Hadoop KMS .................................. SUCCESS [ 48.790 s] [INFO] Apache Hadoop Common Project ....................... SUCCESS [ 0.316 s] [INFO] Apache Hadoop HDFS ................................. SUCCESS [14:58 min] [INFO] Apache Hadoop HttpFS ............................... SUCCESS [11:10 min] [INFO] Apache Hadoop HDFS BookKeeper Journal .............. SUCCESS [01:43 min] [INFO] Apache Hadoop HDFS-NFS ............................. SUCCESS [ 27.438 s] [INFO] Apache Hadoop HDFS Project ......................... SUCCESS [ 0.146 s] [INFO] hadoop-yarn ........................................ SUCCESS [ 0.165 s] [INFO] hadoop-yarn-api .................................... SUCCESS [07:03 min] [INFO] hadoop-yarn-common ................................. SUCCESS [03:31 min] [INFO] hadoop-yarn-server ................................. SUCCESS [ 0.827 s] [INFO] hadoop-yarn-server-common .......................... SUCCESS [01:11 min] [INFO] hadoop-yarn-server-nodemanager ..................... SUCCESS [02:25 min] [INFO] hadoop-yarn-server-web-proxy ....................... SUCCESS [ 17.129 s] [INFO] hadoop-yarn-server-applicationhistoryservice ....... SUCCESS [ 39.350 s] [INFO] hadoop-yarn-server-resourcemanager ................. SUCCESS [01:44 min] [INFO] hadoop-yarn-server-tests ........................... SUCCESS [ 32.941 s] [INFO] hadoop-yarn-client ................................. SUCCESS [ 44.664 s] [INFO] hadoop-yarn-applications ........................... SUCCESS [ 0.197 s] [INFO] hadoop-yarn-applications-distributedshell .......... SUCCESS [ 15.165 s] [INFO] hadoop-yarn-applications-unmanaged-am-launcher ..... SUCCESS [ 9.604 s] [INFO] hadoop-yarn-site ................................... SUCCESS [ 0.149 s] [INFO] hadoop-yarn-registry ............................... SUCCESS [ 31.971 s] [INFO] hadoop-yarn-project ................................ SUCCESS [ 22.195 s] [INFO] hadoop-mapreduce-client ............................ SUCCESS [ 0.673 s] [INFO] hadoop-mapreduce-client-core ....................... SUCCESS [02:08 min] [INFO] hadoop-mapreduce-client-common ..................... SUCCESS [01:38 min] [INFO] hadoop-mapreduce-client-shuffle .................... SUCCESS [ 24.796 s] [INFO] hadoop-mapreduce-client-app ........................ SUCCESS [01:02 min] [INFO] hadoop-mapreduce-client-hs ......................... SUCCESS [ 43.043 s] [INFO] hadoop-mapreduce-client-jobclient .................. SUCCESS [01:09 min] [INFO] hadoop-mapreduce-client-hs-plugins ................. SUCCESS [ 9.662 s] [INFO] Apache Hadoop MapReduce Examples ................... SUCCESS [ 40.439 s] [INFO] hadoop-mapreduce ................................... SUCCESS [ 13.894 s] [INFO] Apache Hadoop MapReduce Streaming .................. SUCCESS [ 32.797 s] [INFO] Apache Hadoop Distributed Copy ..................... SUCCESS [01:00 min] [INFO] Apache Hadoop Archives ............................. SUCCESS [ 11.333 s] [INFO] Apache Hadoop Rumen ................................ SUCCESS [ 35.122 s] [INFO] Apache Hadoop Gridmix .............................. SUCCESS [ 22.939 s] [INFO] Apache Hadoop Data Join ............................ SUCCESS [ 17.568 s] [INFO] Apache Hadoop Ant Tasks ............................ SUCCESS [ 12.339 s] [INFO] Apache Hadoop Extras ............................... SUCCESS [ 18.325 s] [INFO] Apache Hadoop Pipes ................................ SUCCESS [ 27.889 s] [INFO] Apache Hadoop OpenStack support .................... SUCCESS [ 30.148 s] [INFO] Apache Hadoop Amazon Web Services support .......... SUCCESS [01:28 min] [INFO] Apache Hadoop Client ............................... SUCCESS [ 25.086 s] [INFO] Apache Hadoop Mini-Cluster ......................... SUCCESS [ 0.657 s] [INFO] Apache Hadoop Scheduler Load Simulator ............. SUCCESS [ 25.302 s] [INFO] Apache Hadoop Tools Dist ........................... SUCCESS [ 23.268 s] [INFO] Apache Hadoop Tools ................................ SUCCESS [ 0.156 s] [INFO] Apache Hadoop Distribution ......................... SUCCESS [01:06 min] [INFO] ------------------------------------------------------------------------ [INFO] BUILD SUCCESS [INFO] ------------------------------------------------------------------------ [INFO] Total time: 01:17 h [INFO] Finished at: 2014-12-29T20:45:54-08:00 [INFO] Final Memory: 94M/193M [INFO] ------------------------------------------------------------------------ [root@Master hadoop-2.6.0-src]#

二.hadoop + hbase +zookeeper 环境搭建

1.修改Linux主机名 2.修改IP 3.修改主机名和IP的映射关系 ######注意######如果你们公司是租用的服务器或是使用的云主机（如华为用主机、阿里云主机等） /etc/hosts里面要配置的是内网IP地址和主机名的映射关系 4.关闭防火墙 5.ssh免登陆 6.安装JDK，配置环境变量等集群规划：主机名 IP 安装的软件运行的进程 Master 192.168.1.201 jdk、hadoop NameNode、DFSZKFailoverController(zkfc) Slave1 192.168.1.202 jdk、hadoop NameNode、DFSZKFailoverController(zkfc) Slave2 192.168.1.203 jdk、hadoop ResourceManager Slave3 192.168.1.204 jdk、hadoop ResourceManager Slave4 192.168.1.205 jdk、hadoop、zookeeper DataNode、NodeManager、JournalNode、QuorumPeerMain Slave5 192.168.1.206 jdk、hadoop、zookeeper DataNode、NodeManager、JournalNode、QuorumPeerMain Slave6 192.168.1.207 jdk、hadoop、zookeeper DataNode、NodeManager、JournalNode、QuorumPeerMain 复制代码说明：

1.在hadoop2.0中通常由两个NameNode组成，一个处于active状态，另一个处于standby状态。Active NameNode对外提供服务，而Standby NameNode则不对外提供服务，仅同步active namenode的状态，以便能够在它失败时快速进行切换。 hadoop2.0官方提供了两种HDFS HA的解决方案，一种是NFS，另一种是QJM。这里我们使用简单的QJM。在该方案中，主备NameNode之间通过一组JournalNode同步元数据信息，一条数据只要成功写入多数JournalNode即认为写入成功。通常配置奇数个JournalNode 这里还配置了一个zookeeper集群，用于ZKFC（DFSZKFailoverController）故障转移，当Active NameNode挂掉了，会自动切换Standby NameNode为standby状态

2.hadoop-2.2.0中依然存在一个问题，就是ResourceManager只有一个，存在单点故障，hadoop-2.4.1解决了这个问题，有两个ResourceManager，一个是Active，一个是Standby，状态由zookeeper进行协调

安装步骤： 1.安装配置zooekeeper集群（在Slave4上） 1.1解压 [root@Master local]#tar -zxvf zookeeper-3.4.6.tar.g-C /usr/local/ [root@Master local]#mv zookeeper-3.4.6/ zookeeper 复制代码 1.2修改配置 [root@Master local]#cd /usr/local/zookeeper/conf/ [root@Master local]#cp zoo_sample.cfg zoo.cfg [root@Master local]#vim zoo.cfg 复制代码修改： dataDir=/itcast/zookeeper/zkData 复制代码在最后添加： server.1=Slave4:2888:3888 server.2=Slave5:2888:3888 server.3=Slave6:2888:3888 复制代码保存退出然后创建一个tmp文件夹 [root@Master local]#mkdir /usr/local/zookeeper/zkData 复制代码再创建一个空文件 [root@Master local]#touch /usr/local/zookeeper/zkData/myid 复制代码最后向该文件写入ID [root@Master local]#echo 1 > /usr/local/zookeeper/zkData/myid 复制代码 1.3将配置好的zookeeper拷贝到其他节点(首先分别在Slave5、Slave6根目录：/usr/local/) [root@Master local]#scp -r /usr/local/zookeeper/ Slave5:/usr/local/ [root@Master local]#scp -r /usr/local/zookeeper/ Slave6:/usr/local/ 复制代码注意：修改Slave5、Slave6对应/usr/local/zookeeper/zkData/myid内容 Slave5： [root@Master local]#echo 2 > /usr/local/zookeeper/zkData/myid Slave6： [root@Master local]#echo 3 > /usr/local/zookeeper/zkData/myid 复制代码 2.安装配置hadoop集群（在Master上操作） 2.1解压 [root@Master local]#tar -zxvf hadoop-2.6.0.tar.gz -C /usr/local/ 复制代码 2.2配置HDFS（hadoop2.0所有的配置文件都在$HADOOP_HOME/etc/hadoop目录下） #将hadoop添加到环境变量中 [root@Master local]#vim /etc/profile export JAVA_HOME=/usr/local/jdk1.7 export HADOOP_HOME=/usr/local/hadoop-2.6.0 export PATH=$PATH:$JAVA_HOME/bin:$HADOOP_HOME/bin 复制代码 #hadoop2.0的配置文件全部在$HADOOP_HOME/etc/hadoop下 [root@Master local]#cd /usr/local/hadoop-2.6.0/etc/hadoop 复制代码 2.2.1修改hadoo-env.sh export JAVA_HOME=/usr/local/jdk1.7 复制代码 2.2.2修改core-site.xml <configuration>  <property> <name>fs.defaultFS</name> <value>hdfs://masters</value> </property>  <property> <name>hadoop.tmp.dir</name> <value>/usr/local/hadoop-2.6.0/tmp</value> </property>  <property> <name>ha.zookeeper.quorum</name> <value>Slave4:2181,Slave5:2181,Slave6:2181</value> </property> </configuration> 复制代码 2.2.3修改hdfs-site.xml <configuration>  <property> <name>dfs.nameservices</name> <value>masters,ns1,ns2,ns3</value> </property>  <property> <name>dfs.ha.namenodes.masters</name> <value>Master,Slave1</value> </property>  <property> <name>dfs.namenode.rpc-address.masters.Master</name> <value>Master:9000</value> </property>  <property> <name>dfs.namenode.http-address.masters.Master</name> <value>Master:50070</value> </property>  <property> <name>dfs.namenode.rpc-address.masters.Slave1</name> <value>Slave1:9000</value> </property>  <property> <name>dfs.namenode.http-address.masters.Slave1</name> <value>Slave1:50070</value> </property>  <property> <name>dfs.namenode.shared.edits.dir</name> <value>qjournal://Slave4:8485;Slave5:8485;Slave6:8485/masters</value> </property>  <property> <name>dfs.journalnode.edits.dir</name> <value>/usr/local/hadoop-2.6.0/journal</value> </property>  <property> <name>dfs.ha.automatic-failover.enabled</name> <value>true</value> </property>  <property> <name>dfs.client.failover.proxy.provider.masters</name> <value>org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider</value> </property>  <property> <name>dfs.ha.fencing.methods</name> <value> sshfence shell(/bin/true) </value> </property>  <property> <name>dfs.ha.fencing.ssh.private-key-files</name> <value>/root/.ssh/id_rsa</value> </property>  <property> <name>dfs.ha.fencing.ssh.connect-timeout</name> <value>30000</value> </property> </configuration> 复制代码 2.2.4修改mapred-site.xml <configuration>  <property> <name>mapreduce.framework.name</name> <value>yarn</value> </property> </configuration> 复制代码 2.2.5修改yarn-site.xml <configuration>  <property> <name>yarn.resourcemanager.ha.enabled</name> <value>true</value> </property>  <property> <name>yarn.resourcemanager.cluster-id</name> <value>RM_HA_ID</value> </property>  <property> <name>yarn.resourcemanager.ha.rm-ids</name> <value>rm1,rm2</value> </property>  <property> <name>yarn.resourcemanager.hostname.rm1</name> <value>Slave2</value> </property> <property> <name>yarn.resourcemanager.hostname.rm2</name> <value>Slave3</value> </property> <property> <name>yarn.resourcemanager.recovery.enabled</name> <value>true</value> </property> <property> <name>yarn.resourcemanager.store.class</name> <value>org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore</value> </property>  <property> <name>yarn.resourcemanager.zk-address</name> <value>Slave4:2181,Slave5:2181,Slave6:2181</value> </property> <property> <name>yarn.nodemanager.aux-services</name> <value>mapreduce_shuffle</value> </property> </configuration> 复制代码 2.2.6修改slaves(slaves是指定子节点的位置，因为要在Master上启动HDFS、在Slave2启动yarn，所以Master上的slaves文件指定的是datanode的位置，slave2上的slaves文件指定的是nodemanager的位置) Slave4 Slave5 Slave6 复制代码 2.2.7配置免密码登陆 #首先要配置Master到Slave1、Slave2、Slave3、Slave4、Slave5、Slave6的免密码登陆 #在Master上生产一对钥匙 [root@Master local]#ssh-keygen -t rsa 复制代码 #将公钥拷贝到其他节点，包括自己 [root@Master local]#ssh-copy-id Master [root@Master local]#ssh-copy-id Slave1 [root@Master local]#ssh-copy-id Slave2 [root@Master local]#ssh-copy-id Slave3 [root@Master local]#ssh-copy-id Slave4 [root@Master local]#ssh-copy-id Slave5 [root@Master local]#ssh-copy-id Slave6 复制代码 #配置Slave2到Slave3、Slave4、Slave5、Slave6的免密码登陆 #在Slave2上生产一对钥匙 [root@Master local]#ssh-keygen -t rsa 复制代码 #将公钥拷贝到其他节点 [root@Master local]#ssh-copy-id Slave3 [root@Master local]#ssh-copy-id Slave4 [root@Master local]#ssh-copy-id Slave5 [root@Master local]#ssh-copy-id Slave6 复制代码 #注意：两个namenode之间要配置ssh免密码登陆，别忘了配置Slave1到Master的免登陆在Slave1上生产一对钥匙 [root@Master local]#ssh-keygen -t rsa [root@Master local]#ssh-copy-id -i Master 复制代码 #在Slave3上生产一对钥匙 [root@Master local]#ssh-keygen -t rsa 复制代码 #将公钥拷贝到其他节点 [root@Master local]#ssh-copy-id Slave4 [root@Master local]#ssh-copy-id Slave5 [root@Master local]#ssh-copy-id Slave6 复制代码 2.4将配置好的hadoop拷贝到其他节点 [root@Master local]#scp -r /usr/local/hadoop-2.6.0/ Slave1:/usr/local/ [root@Master local]#scp -r /usr/local/hadoop-2.6.0/ Slave2:/usr/local/ [root@Master local]#scp -r /usr/local/hadoop-2.6.0/ Slave3:/usr/local/ [root@Master local]#scp -r /usr/local/hadoop-2.6.0/ Slave4:/usr/local/ [root@Master local]#scp -r /usr/local/hadoop-2.6.0/ Slave5:/usr/local/ [root@Master local]#scp -r /usr/local/hadoop-2.6.0/ Slave6:/usr/local/ 复制代码 ###注意：严格按照下面的步骤 2.5启动zookeeper集群（分别在Slave4、Slave5、Slave6上启动zk） [root@Master local]#cd /usr/local/zookeeper/bin/ [root@Master local]#./zkServer.sh start 复制代码 #查看状态：一个leader，两个follower [root@Master local]#./zkServer.sh status 复制代码 2.6启动journalnode（分别在Slave4、Slave5、Slave6上执行） [root@Master local]#cd /usr/local/hadoop-2.6.0/sbin [root@Master local]#sbin/hadoop-daemon.sh start journalnode 复制代码 #运行jps命令检验，Slave4、Slave5、Slave6上多了JournalNode进程 2.7格式化HDFS #在Master上执行命令: [root@Master local]#hdfs namenode -format 复制代码 #格式化后会在根据core-site.xml中的hadoop.tmp.dir配置生成个文件，这里我配置的是/usr/local/hadoop-2.6.0/tmp，然后将/usr/local/hadoop-2.6.0/tmp拷贝到Slave1的/usr/local/hadoop-2.6.0/下。 [root@Master local]#scp -r tmp/ Slave1:/usr/local/hadoop-2.6.0/ 复制代码 2.8格式化ZK(在Master上执行即可) [root@Master local]#hdfs zkfc -formatZK 复制代码 2.9启动HDFS(在Master上执行) [root@Master local]#sbin/start-dfs.sh 复制代码 2.10启动YARN(#####注意#####：是在Slave2上执行start-yarn.sh，把namenode和resourcemanager分开是因为性能问题，因为他们都要占用大量资源，所以把他们分开了，他们分开了就要分别在不同的机器上启动) [root@Master local]#Slave2:${HADOOP_HOME}/sbin/start-yarn.sh [root@Master local]#Slave3:${HADOOP_HOME}/sbin/yarn-daemon.sh start resourcemanager 复制代码到此，hadoop-2.6.0配置完毕，可以统计浏览器访问: http://192.168.80.100:50070 NameNode 'Master:9000' (active) http://192.168.80.101:50070 NameNode 'Slave1:9000' (standby) 复制代码验证HDFS HA 首先向hdfs上传一个文件 [root@Master local]#hadoop fs -put /etc/profile /profile [root@Master local]#hadoop fs -ls / 复制代码然后再kill掉active的NameNode [root@Master local]#kill -9 <pid of NN> 复制代码通过浏览器访问： http://192.168.80.101:50070 NameNode 'Slave1:9000' (active) 这个时候Slave1上的NameNode变成了active 在执行命令： [root@Master local]#hadoop fs -ls / -rw-r--r-- 3 root supergroup 1926 2014-02-06 15:36 /profile 复制代码刚才上传的文件依然存在！！！手动启动那个挂掉的NameNode [root@Master local]#sbin/hadoop-daemon.sh start namenode 复制代码通过浏览器访问： http://192.168.80.101:50070 NameNode 'Master:9000' (standby) 复制代码验证YARN：运行一下hadoop提供的demo中的WordCount程序： [root@Master local]#hadoop jar share/hadoop/mapreduce/hadoop-mapreduce-examples-2.4.1.jar wordcount /profile /out 复制代码 hadoop HA集群搭建完成 hbase-0.98.9-hadoop2 搭建 4.1 解压缩，并重命名 [root@Master local]#mv hbase-** hbase 复制代码修改环境变量： export HBASE_HOME=/usr/local/hbase export PATH= .:$PATH:$HBASE_HOME/bin: 复制代码保存，退出。执行 source /etc/profile 生效 4.1 修改HBase的配置文件#HBASE_HOME/conf/hbase-env.sh 修改内容如下： export JAVA_HOME=usr/local/jdk/ export HBASE_MANAGES_ZK=true //HBase是否管理它自己的ZooKeeper的实例。复制代码保存，退出。 4.2 修改HBase的配置文件#HBASE_HOME/conf/hbase-site.xml，修改内容如下： <property> <name>hbase.rootdir</name> <value>hdfs://Master:9000/hbase</value> </property> <property> <name>hbase.cluster.distributed</name> <value>true</value> </property> <property> <name>hbase.zookeeper.quorum</name> <value>Master</value> </property> <property> <name>dfs.replication</name> <value>3</value> </property> 复制代码注意：$HBASE_HOME/conf/hbase-site.xml的hbase.rootdir的主机和端口号与$HADOOP_HOME/conf/core-site.xml的fs.default.name的主机和端口号一致 4.3 (可选)文件 regionservers 的内容修改为Master. 4.4 执行目录到../bin ，执行命令 start-hbase.sh ******启动hbase之前，确保hadoop是运行正常的。并且可以写入文件。 4.5 验证：(1)执行jps ，发现新增加了3个Havana进程，分别是HMaster、HRegionServer、HQuorumPeer (HQuorumPeerMain 是ZooKeeper的进程 ) 备注:启动HBase时，请先执行 /usr/local/zookeeper/bin zkServer.sh stop 停止ZooKeeper的进程,以免hbase启动失败。 (2)通过浏览器查看： http://masters:60010 5.HBase的集群安装(在原来的Master上的hbase伪分布基础上搭建)： 5.1 集群结构，主节点(hmaster)是Master，从节点(region server)是Slave1,Slave2,Slave3. 5.2 修改hadoop0上的hbase的几个文件 (1)修改hbase-env.sh 最后一行 export HBASE_MANAGES_ZK=false. (2)修改hbase-site.xml文件的 hbase.zookeeper.quorum 的值为 Master,Slave1,Slave2,Slave3 。 (3)修改regionservers文件(存放的 region server的hostname ),内容修改成 Slave1,Slave2,Slave3 。 5.3 复制Master中的hbase到Slave1,Slave2,Slave3的对应目录下,并复制、Master 的/etc/profile文件到hadoop1 、hadoop2 中。 [root@Master local]#scp -r hbase Slave1:/usr/local/ [root@Master local]#scp -r /etc/profile Slave1:/etc/profile [root@Master local]#source /etc/profile 复制代码 5.4 在HA集群中，首先各个节点启动ZooKeeper集群，其次 Master中启动hadoop集群，最后在Master上启动hbase集群。 6.测试Hbase是否启动正常： 1) 在Master主机中执行jps，查看进程。会新增一个 HMaster 进程 2) 在regionserver 中执行 jps，新增 HRegionServer。 7.执行hbase脚本命令： [root@Slave2 local]# hbase shell 复制代码

最新回复(0)