10、linux上安装hive

1.基本信息

版本	1.2.1
安装机器	三台机器
账号	hadoop
源路径	/opt/software/apache-hive-1.2.1-bin.tar.gz
目标路径	/opt/hive -> opt/apache-hive-1.2.1-bin/
依赖关系	mysql

2.安装过程

1).登陆172.24.5.173，并登陆mysql，创建hive元数据库及hive用户,并设置字符集为latin1：

[root@bgs-5p173-wangwenting ~]#  mysql -u root -p

mysql>create database hive_metastore;

mysql>GRANT ALL PRIVILEGES ON hive_metastore.* TO 'hive'@'%' IDENTIFIED BY 'hive';

mysql>GRANT ALL PRIVILEGES ON hive_metastore.* TO 'hive'@bgs-5p173-wangwenting IDENTIFIED BY 'hive';

mysql>flush privileges;

mysql>alter database hive_metastore character set latin1;

2).创建hadoop目录，并修改权限：

[root@bgs-5p173-wangwenting ~]# su - hadoop

[hadoop@bgs-5p173-wangwenting ~]$ hadoop fs -mkdir -p /user/hive/warehouse

[hadoop@bgs-5p173-wangwenting ~]$ hadoop fs -chmod 777 /tmp

[hadoop@bgs-5p173-wangwenting ~]$ hadoop fs -chmod 777 /user/hive/warehouse

3).将apache-hive-${version}-bin.tar.gz通过tar -zxvf命令解压缩至目的安装目录,并设置软连接:

[hadoop@bgs-5p173-wangwenting ~]$ cd /opt/software

[hadoop@bgs-5p173-wangwenting software]$ tar -zxvf apache-hive-${version}-bin.tar.gz -C /opt

[hadoop@bgs-5p173-wangwenting software]$ cd /opt

[hadoop@bgs-5p173-wangwenting opt]$ ln -s /opt/apache-hive-${version}-bin   /opt/hive

4).删除hadoop目录中的jline-0.9.94.jar,防止jar冲突（若存在）:

[hadoop@bgs-5p173-wangwenting hive]$  rm -r /opt/hadoop/share/hadoop/yarn/lib/jline-0.9.94.jar

在/opt/hadoop/hive/conf目录下添加hive-site.xml,并修改hive-site.xml配置文件:

[hadoop@bgs-5p173-wangwenting hive]$ cd /opt/hive/conf

[hadoop@bgs-5p173-wangwenting conf]$ cp hive-default.xml.template hive-site.xml

[hadoop@bgs-5p173-wangwenting conf]$ vim hive-site.xml

添加下面内容:（可以将里面的内容全部删除只添加以下内容，我的是追加的形式，不过要修改里面的配置内容。${system-user.name}的值改为hive,${java.io.tmpdir }改为/tmp）

<configuration>

 <property>

 <name>javax.jdo.option.ConnectionURL</name>

 <value>jdbc:mysql://bgs-5p173-wangwenting:3306/hive_metastore?createDatabaseIfNotExist=true</value>

 </property>

 <property>

 <name>javax.jdo.option.ConnectionDriverName</name>

 <value>com.mysql.jdbc.Driver</value>

 </property>

 <property>

 <name>javax.jdo.option.ConnectionUserName</name>

 <value>hive</value>

 </property>

 <property>

 <name>javax.jdo.option.ConnectionPassword</name>

 <value>hive</value>

 </property>

 <property>

 <name>hive.metastore.warehouse.dir</name>

 <value>/user/hive/warehouse</value>

 </property>

 <property>

 <name>hive.querylog.location</name>

 <value>/tmp/hive</value>

 </property>

 <property>

 <name>hive.server2.logging.operation.log.location</name>

 <value>/tmp/hive/operation_logs</value>

 </property>

 <property>

 <name>hive.exec.scratchdir</name>

 <value>/tmp/hive</value>

 </property>

</configuration>

6).确保启动hadoop集群的情况下，切换到hadoop账户

[hadoop@bgs-5p173-wangwenting conf]$ nohup /opt/hive/bin/hive &

7).启动hiveserver2：

[hadoop@bgs-5p173-wangwenting conf]$ nohup /opt/hive/bin/hive --service hiveserver2 &

使用beeline工具连接到thrift server

[hadoop@bgs-5p173-wangwenting conf]$ /opt/hive/bin/beeline -u jdbc:hive2://bgs-5p173-wangwenting:10000

如果连接失败需要拷贝一个jar包到lib目录下mysql-connector-java-5.1.35-bin.jar

8).测试hive基本命令是否正常运行:

hive> create table test(a int);

OK

Time taken: 0.55 seconds

hive> show tables;

OK

Test

Time taken: 0.031 seconds, Fetched: 1 row(s)

hive> desc test;

OK

a                       int                                        

Time taken: 0.157 seconds, Fetched: 1 row(s)

hive> select a from test;

OK

Time taken: 0.14 seconds

hive> drop table test;

OK

Time taken: 0.671 seconds

hive> show tables;

OK

Time taken: 0.032 seconds

如果测试同上，则说明hive安装成功，能够正常运行。

如果报错，需要修改conf目录下的hive-site.xml文件，将${system:user.name} 换成hive即可