[原创]独立模式安装Hive

 

独立模式安装Hive

 

在三台CentOS虚拟机server1、server2、 server3上已经成功安装全分布模式的Hadoop集群，其中server1为NameNode, server2、server3为DataNode。本人在server1即NameNode上安装Hive，并使用MySQL数据库存放metadata。

 

1.首先配置MySQL

安装MySQL

[root@server1 ~]# yum install mysql mysql -server

一路选择yes就行

 

添加MySQL服务

[root@server1 ~]# /sbin/chkconfig --add mysqld

 

启动MySQL

[root@server1 ~]# service mysqld start

Starting mysqld:                                           [  OK  ]

 

用root账户在本地登录MySQL

[root@server1 ~]# mysql -u root

出现欢迎页面，进入MySQL monitor.

 

创建数据库实例hive

mysql > CREATE DATABASE hive;

 

创建用户hive

mysql > CREATE USER 'hive' IDENTIFIED BY 'hive';

 

给用户hive赋予相应的访问与读写权限

mysql > GRANT ALL ON hive.* TO hive@localhost ;

 

2.配置Hive

 

下载hive-0.8.1，并解压

[admin@server1 hive-0.8.1]$ pwd

/home/admin/hive-0.8.1

 

修改系统环境变量，在path中加入Hive执行路径

[root@server1 ~]# vim /etc/profile

...................

export HIVE_HOME=/home/admin/hive-0.8.1

export PATH=$PATH:$HIVE_HOME/bin

应用修改

[root@server1 ~]# source /etc/profile

 

下载MySQL的JDBC驱动包MySQL-connector-java-5.1.18-bin.jar，复制到hive-0.8.1的lib目录下

 

将Hive的conf下的文件hive-default.xml.template复制一份，重命名为hive-site.xml

 

[admin@server1 conf]$ cp hive-default.xml.template hive-site.xml

 

将复制的hive-site.xml 移动到所安装的Hadoop的conf目录下

[admin@server1 conf]$ mv hive-site.xml  /home/admin/hadoop-0.20.2/conf

 

修改hive-site.xml，配置，主要修改以下属性

 

所连接的MySQL数据库实例

 <property>

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

  <value>jdbc:mysql://localhost:3306/hive</value>

  <description>JDBC connect string for a JDBC metastore</description>

</property>

 

连接的MySQL数据库驱动

<property>

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

  <value>com.mysql.jdbc.Driver</value>

  <description>Driver class name for a JDBC metastore</description>

</property>

 

连接的MySQL数据库用户名

<property>

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

  <value>hive</value>

  <description>username to use against metastore database</description>

</property>

 

连接的MySQL数据库密码

<property>

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

  <value>hive</value>

  <description>password to use against metastore database</description>

</property>

 

2.测试Hive

在控制台键入Hive，进入Hive操作界面

 

[admin@server1 ~]$ hive

Logging initialized using configuration in jar:file:/home/admin/hive-0.8.1/lib/hive-common-0.8.1.jar!/hive-log4j.properties

Hive history file=/tmp/admin/hive_job_log_admin_201212011113_1138680566.txt

hive>

 

显示存在的表：

hive> SHOW TABLES;      

OK

Time taken: 0.063 seconds

 

建立表records：

hive> CREATE TABLE records (year STRING, temperature INT, quality INT)

    > ROW FORMAT DELIMITED FIELDS TERMINATED BY '\t';               

OK

Time taken: 0.253 seconds

 

显示存在的表，多了一个records：

hive> SHOW TABLES;                                                  

OK

records

Time taken: 0.089 seconds

 

查看表records的定义

hive> DESCRIBE records;

OK

year string     

temperature    int  

quality    int  

Time taken: 0.139 seconds

 

向表records导入数据

hive> LOAD DATA INPATH '/user/admin/in/ncdc/micro/sample.txt'

    > INTO TABLE records;                                    

Loading data to table default.records

OK

Time taken: 0.337 seconds

 

查看表records的数据

hive> SELECT * FROM records;

OK

1950       0     1

1950       22    1

1950       -11   1

1949       111  1

1949       78    1

Time taken: 0.264 seconds

 

计算records中每一年的最高温度

hive> SELECT year, MAX(temperature) FROM records GROUP BY year;        

Total MapReduce jobs = 1

Launching Job 1 out of 1

Number of reduce tasks not specified. Estimated from input data size: 1

In order to change the average load for a reducer (in bytes):

  set hive.exec.reducers.bytes.per.reducer=<number>

In order to limit the maximum number of reducers:

  set hive.exec.reducers.max=<number>

In order to set a constant number of reducers:

  set mapred.reduce.tasks=<number>

Starting Job = job_201211240040_0037, Tracking URL = http://server1:50030/jobdetails.jsp?jobid=job_201211240040_0037

Kill Command = /home/admin/hadoop-0.20.2/bin/../bin/hadoop job  -Dmapred.job.tracker=server1:9001 -kill job_201211240040_0037

Hadoop job information for Stage-1: number of mappers: 1; number of reducers: 1

2012-12-01 11:30:57,089 Stage-1 map = 0%,  reduce = 0%

2012-12-01 11:31:15,188 Stage-1 map = 100%,  reduce = 0%

2012-12-01 11:31:24,235 Stage-1 map = 100%,  reduce = 100%

Ended Job = job_201211240040_0037

MapReduce Jobs Launched:

Job 0: Map: 1  Reduce: 1   HDFS Read: 51 HDFS Write: 17 SUCESS

Total MapReduce CPU Time Spent: 0 msec

OK

1949       111

1950       22

Time taken: 47.238 seconds