Summary:
hadoop stand-alone, pseudo-distributed, distributed installation
hadoop cluster complete summary of the issues encountered
Distributed Hbase complete installation steps:
note. I hbase cluster is to continue to build from in front of the hadoop cluster (after the input nodes to build a good jps each node to start the process as follows :)
Host computer | Aliases | install software | Existing processes | Service Address |
192.168.248.138 | cdh1 | hadoop2.8 jdk1.8 | namenode DFSZKFailoverController HMaster | http://cdh1:50070 http://cdh1:16010/master-status |
192.168.248.139 | cdh2 | hadoop2.8 jdk1.8 | namenode DFSZKFailoverController | http://cdh1:50070 |
192.168.248.140 | cdh3 | hadoop2.8 jdk1.8 | ResourceManager | |
192.168.248.141 | Cdh4 | hadoop2.8 jdk1.8 zookeeper3.4.13 | QuorumPeerMain JournalNode DataNode NodeManager | |
192.168.248.142 | cdh5 | hadoop2.8 jdk1.8 zookeeper3.4.13 hbase1.4.0 | QuorumPeerMain JournalNode DataNode NodeManager HMaster HRegionServer | http://cdh5:16010/master-status |
192.168.248.143 | cdh6 | hadoop2.8 jdk1.8 zookeeper3.4.13 hbase1.4.0 | QuorumPeerMain JournalNode DataNode NodeManager HRegionServer | |
192.168.248.144 | cdh7 | hadoop2.8 jdk1.8 hbase1.4.0 | JournalNode DataNode NodeManager HRegionServer |
1> select a compatible version hadoop2.8 hbase, and in this I chose hbase1.4. Specific versions corresponding reference Baidu
2> uploaded to / hadoop folder, extract (root privileges)
3>, and arranged hbase-env.sh profile hbase-site.xml
Note that before modifying the configuration files, environment variables hbase first configuration to / etc / profile, this step build hadoop cluster have already said many times, in this direct screenshot
Modified into hbase-env.sh j / HBASE / conf under
export JAVA_HOME = / hadoop / jdk1.8.0_181 revise their jdk
export HBASE_MANAGES_ZK = false without hbase comes zk, zk use our cluster
Modify hbase-site.xml: [a closer look at their own will understand]
4> from the node configuration modification regionservers
Note: this I should start on cdh5, so cdh5 is HMaster, for high availability, I was on namedata node cdh1 start, so cdh1 and cdh5 are Hmater, while cdh5, cdh6, cdh7 completed structures like this are HRegionServer [ we will see]
5> hbase synchronized to the machine cdh5 cdh6 cdh7
scp -r $HBASE_HOME cdh5:/hadoop/
scp -r $HBASE_HOME cdh6:/hadoop/
scp -r $HBASE_HOME cdh7:/hadoop/
6> starts on the primary hbase on cdh1, along with the entire cluster will be started
Note: Note: Note: Before starting hbase cluster must ensure time synchronization, or not start hbase cluster, the cluster can be started or part of the time range, remember
时间同步最简的是 date -s "2019-05-31 09:02:00" 【这种方式重启后时间同步失效】 还可以用netdate (这种重启后不会失效,采用在线同步的方式同步时间)
时间同步完成后就可以启动hbase集群 进入$HBASE_HOME/bin 执行 start-hbase-sh 即可
到这hbase完整分布式集群也搭建完成了。
Web访问:
http://192.168.248.138:16010/master-status