1.版本(当前实验版本组合,版本一定要兼容,否则各种问题)
Hadoop2.7.2+HBase1.1.5+Hive2.0.0
kylin-1.5.2.1 (apache-kylin-1.5.1-HBase1.1.3-bin.tar.gz)
2.Hadoop环境编译以支持Snappy解压缩库
重新编译hadoop-2.7.2-src的native以支持Snappy解压压缩库
3.环境准备
Hadoop-2.7.2+zookeeper-3.4.6完全分布式环境搭建(HDFS、YARN HA)
Hadoop2.7实战v1.0之HBase1.1.5 HA分布式搭建
配置Hadoop2.7.2和Hbase1.1.5支持Snappy解压压缩库
Hadoop2.7实战v1.0之Hive-2.0.0+MySQL本地模式安装
4. 配置Hadoop2.7.2和Hbase1.1.5支持Snappy解压压缩库
5. 主机规划
IP Host 安装软件 进程
172.16.101.55 sht-sgmhadoopnn-01 apache-kylin-1.5.1-HBase1.1.3-bin.tar.gz RunJar
6. 添加hive_dependency和KYLIN_HOME环境变量
点击(此处)折叠或打开
- [root@sht-sgmhadoopnn-01 ~]# vi /etc/profile
- export JAVA_HOME=/usr/java/jdk1.7.0_67-cloudera
- export CLASSPATH=.:$JAVA_HOME/jre/lib/rt.jar:$JAVA_HOME/lib/dt.jar:$JAVA_HOME/lib/tools.jar
-
- export HADOOP_HOME=/hadoop/hadoop
- export HBASE_HOME=/hadoop/hbase
- export ZOOKEEPER_HOME=/hadoop/zookeeper
-
- export HIVE_HOME=/hadoop/hive
- export hive_dependency=/hadoop/hive/conf:/hadoop/hive/lib/*:/hadoop/hive/hcatalog/share/hcatalog/hive-hcatalog-core-2.0.0.jar
-
- export KYLIN_HOME=/hadoop/kylin
-
- export PATH=.:$HADOOP_HOME/bin:$JAVA_HOME/bin:$ZOOKEEPER_HOME/bin:$HBASE_HOME/bin:$HIVE_HOME/bin:$KYLIN_HOME/bin:$PATH
-
- "/etc/profile" 99L, 2592C written
- [root@sht-sgmhadoopnn-01 ~]# source /etc/profile
7.同步/etc/profile到集群其他节点,生效
8.下载安装配置kylin
点击(此处)折叠或打开
- # 在sht-sgmhadoopnn-01上执行以下命令
- [root@sht-sgmhadoopnn-01 ~]# cd /hadoop
- [root@sht-sgmhadoopnn-01 hadoop]# wget https://dist.apache.org/repos/dist/release/kylin/apache-kylin-1.5.2.1/apache-kylin-1.5.2.1-HBase1.x-bin.tar.gz
- [root@sht-sgmhadoopnn-01 hadoop]# tar -zxvf apache-kylin-1.5.1-HBase1.1.3-bin.tar.gz
- [root@sht-sgmhadoopnn-01 hadoop]# ln -s /hadoop/apache-kylin-1.5.2.1-bin /hadoop/kylin
-
- [root@sht-sgmhadoopnn-01 hadoop]# vi /hadoop/kylin/bin/kylin.sh
- export KYLIN_HOME=/hadoop/kylin # 改成绝对路径
- export HBASE_CLASSPATH_PREFIX=${tomcat_root}/bin/bootstrap.jar:${tomcat_root}/bin/tomcat-juli.jar:${tomcat_root}/lib/*:$hive_dependency:$HBASE_CLASSPATH_PREFIX #在路径中添加$hive_dependency
9.修改kylin.properties
点击(此处)折叠或打开
######修改
# List of web servers in use, this enables one web server instance to sync up with other servers.
kylin.rest.servers=sht-sgmhadoopnn-01:7070
######新增
kylin.job.jar=$KYLIN_HOME/lib/kylin-job-1.5.2.1.jar
kylin.coprocessor.local.jar=$KYLIN_HOME /lib/kylin-coprocessor-1.5.2.1.jar
[root@sht-sgmhadoopnn-01 kylin]# $KYLIN_HOME/bin/kylin.sh start
.........
.........
KYLIN_DEBUG_SETTINGS is not set, will not enable remote debuging
KYLIN_LD_LIBRARY_SETTINGS is not set, Usually it's okay unless you want to specify your own native path
A new Kylin instance is started by root, stop it using "kylin.sh stop"
Please visit <a href="http://http://:7070/kylin
You can check the log at /hadoop/kylin/logs/kylin.log<a href="http://http://:7070/kylin
You can check the log at
###http://172.16.101.55:7070/kylin 账号: ADMIN 密码:KYLIN
11.检查kylin是否启动成功
点击(此处)折叠或打开
tcp 0 0 0.0.0.0:7070 0.0.0.0:* LISTEN 30939/java
[root@sht-sgmhadoopnn-01 kylin]#
12.导入官网测试案例
点击(此处)折叠或打开
[root@sht-sgmhadoopnn-01 kylin]# $KYLIN_HOME/bin/sample.sh
[root@sht-sgmhadoopnn-01 kylin]# $KYLIN_HOME/bin/kylin.sh start
1>.选中'kylin_sales_cube'示例立方体，点击'Actions'->'Build'，选择一个截止日期，本试验中选择的是'2016-01-10'
(具体小时,分,秒随便选,因为最终hive sql的语句类似 WHERE (KYLIN_SALES.PART_DT >= '2012-01-01' AND KYLIN_SALES.PART_DT < '2016-01-10');)
2>.在'Monitor'标签中通过刷新页面检查进度条，直到100%
3>.在'Insight'标签中执行下面的SQL查询：
select part_dt, sum(price) as total_selled, count(distinct seller_id) as sellers from kylin_sales group by part_dt order by part_dt;
###耗时2.87s
4>.在hive中执行同一个SQL查询，验证kylin的查询结果(会开启MapReduce Job计算)
###耗时65.205s
查询执行和结果如图所示
Build成功后，hive中建立了3+n个表，如图所示(3个官网案例hive表,n个build的hive表)
Build成功后，hbase中建立了1+n个表，如图所示(1个元数据表,n个build的hbase表)
You can check the log at
