1.hbase is a distributed system, which needs to rely on HDFS as the storage medium and zookeeper to complete the monitoring of the master-slave nodes.
2. Installation preparation is as follows:
1.HBase role HMaster (master node, one Active, one StandBy) Hreginserver (data node, multiple) 2. dependence HDFS cluster, zookeeper cluster [start]
3. Upload the installation package and decompress it
tar -zxvf hbase-1.2.1.bin.tar.gz -C apps/
4. Enter the conf directory to modify the configuration file as follows
1.vi hbase-env.sh export JAVA_HOME=/usr/local/java/jdk1.8.0_231 export HBASE_MANAGES_ZK=false #Own zookeeper, do not start its built-in 2.vi hbase-site.xml <configuration> <!-- Appoint hbase stay HDFS Path stored on --> <property> <name>hbase.rootdir</name> <value>hdfs://hadoop01:9000/hbase</value> </property> <!-- Appoint hbase It's distributed --> <property> <name>hbase.cluster.distributed</name> <value>true</value> </property> <!-- Appoint zk Address, multiple“,"Division --> <property> <name>hbase.zookeeper.quorum</name> <value>hadoop01:2181,hadoop02:2181,hadoop03:2181</value> </property> <!-- Appoint hbase Monitoring page port of --> <property> <name>hbase.master.info.port</name> <value>60010</value> </property> </configuration> 3.vi regionservers hadoop01 hadoop02 hadoop03
5. Copy to other machines
scp -r hbase-1.2.1/ hadoop02:$PWD scp -r hbase-1.2.1/ hadoop03:$PWD
6. Check whether hdfs and zookeeper are started before starting
# 1. Start one by one bin/hbase-daemon.sh start master bin/hbase-daemon.sh start regionserver # 2. Start it all and remember the same time. An error will occur if you do not synchronize. bin/start-hbase.sh # 3. To enable the standby master, you only need to execute it on another machine bin/hbase-daemon.sh start master # 3. Hadoop 01:16010 you can view the web page
7. The command-line client experience is enough, which is usually accessed by java program
cd bin/hbase bin/hbase shell >status >Building watches and so on...
8.java client
slightly