实验三:HBase组件安装与配置
实验目的
完成本实验,您应该能够:
- 掌握HBase安装与配置
- 掌握HBase常用Shell命令
实验要求
- 了解HBase原理
- 熟悉HBase常用Shell命令
实验环境
本实验所需之主要资源环境如表1-1所示。
表1-1 资源环境
服务器****集群 | 3个以上节点,节点间网络互通,各节点最低配置:双核CPU、8GB内存、100G硬盘 |
---|---|
运行环境 | CentOS 7.4 |
大数据平台 | H3C教学与实践管理系统 |
服务和组件 | 完成前面章节实验,其他服务及组件根据实验需求安装 |
实验过程
实验任务一:HBase安装与配置
步骤一:解压缩HBase安装包
[root@master ~]# tar -zxvf /opt/software/hbase-1.2.1-bin.tar.gz -C /usr/local/src/
步骤二:重命名HBase安装文件夹
[root@master ~]# cd /usr/local/src/
[root@master src]# mv hbase-1.2.1 hbase
步骤三:在所有节点添加环境变量
[root@master src]# vi /etc/profile
# set hbase environment
export HBASE_HOME=/usr/local/src/hbase
export PATH=$HBASE_HOME/bin:$PATH
[root@master src]# scp /etc/profile root@slave1:/etc/
root@slave1's password:
profile 100% 2366 2.3MB/s 00:00
[root@master src]# scp /etc/profile root@slave2:/etc/
root@slave2's password:
profile 100% 2366 903.9KB/s 00:00
步骤四:在master节点进入配置文件目录
[root@master ~]# cd /usr/local/src/hbase/conf/
步骤五:在master节点配置hbase-env.sh文件
#在文件中修改
[root @master conf]# vi hbase-env.sh
#Java安装位置
export JAVA_HOME=/usr/local/src/java
#值为true使用HBase自带的ZooKeeper,值为false使用在Hadoop上装的ZooKeeper
export HBASE_MANAGES_ZK=false
#HBase类路径
export HBASE_CLASSPATH=/usr/local/src/hadoop/etc/hadoop/
步骤六:在master节点配置hbase-site.xml
[root @master conf]# vi hbase-site.xml
<property>
<name>hbase.rootdir</name>
<value>hdfs://master:9000/hbase</value> # 使用9000端口
<description>The directory shared by region servers.</description>
</property>
<property>
<name>hbase.master.info.port</name>
<value>60010</value> # 使用master节点60010端口
</property>
<property>
<name>hbase.zookeeper.property.clientPort</name>
<value>2181</value> # 使用master节点2181端口
<description>Property from ZooKeeper's config zoo.cfg. The port at which the clients will connect.
</description>
</property>
<property>
<name>zookeeper.session.timeout</name>
<value>120000</value> # ZooKeeper超时时间
</property>
<property>
<name>hbase.zookeeper.quorum</name>
<value>master,slave1,slave2</value> # ZooKeeper管理节点
</property>
<property>
<name>hbase.tmp.dir</name>
<value>/usr/local/src/hbase/tmp</value> # HBase临时文件路径
</property>
<property>
<name>hbase.cluster.distributed</name>
<value>true</value> # 使用分布式HBase
</property>
hbase.rootdir:该项配置了数据写入的目录,默认hbase.rootdir是指向/tmp/hbase-${user.name},也就说你会在重启后丢失数据(重启的时候操作系统会清理/tmp目录)。
hbase.zookeeper.property.clientPort:指定zk的连接端口
zookeeper.session.timeout:RegionServer与ZooKeeper间的连接超时时间。当超时时间到后,ReigonServer会被ZooKeeper从RS集群清单中移除,HMaster收到移除通知后,会对这台server负责的regions重新balance,让其他存活的RegionServer接管.
hbase.zookeeper.quorum:默认值是localhost,列出zookeepr
ensemble中的servers
hbase.master.info.port:浏览器的访问端口
步骤七:在master节点修改regionservers文件
#删除localhost,每一行写一个slave节点主机机器名
[root @master conf]$ vi regionservers
slave1
slave2
步骤八:在master节点创建hbase.tmp.dir目录
[root @master usr]# mkdir /usr/local/src/hbase/tmp
步骤九:将master上的hbase安装文件同步到slave1 slave2
[root @master ~]# scp -r /usr/local/src/hbase/ root@slave1:/usr/local/src/
[root @master ~]# scp -r /usr/local/src/hbase/ root@slave2:/usr/local/src/
步骤十:在所有节点修改hbase目录权限
master节点修改权限
[root @master ~]# chown -R hadoop:hadoop /usr/local/src/hbase/
slave1节点修改权限
[root @slave1 ~]# chown -R hadoop:hadoop /usr/local/src/hbase/
slave2节点修改权限
[root @slave2 ~]# chown -R hadoop:hadoop /usr/local/src/hbase/
步骤十一:在所有节点切换到hadoop用户并使环境变量生效
master节点
[root @master ~]# su - hadoop
[root @master ~]# source /etc/profile
slave1节点
[root @slave1 ~]# su - hadoop
[root @slave1 ~]# source /etc/profile
slave2节点
[root @slave2~]# su - hadoop
[root @slave2~]# source /etc/profile
步骤十二:启动HBase
先启动Hadoop,然后启动ZooKeeper,最后启动HBase。
首先在master节点启动Hadoop。
[hadoop@master ~]$ start-all.sh
[hadoop@master ~]$ jps
master节点
[hadoop@master ~]$ jps
10288 ResourceManager
9939 NameNode
10547 Jps
10136 SecondaryNameNode
slave1节点
[hadoop@slave1 ~]$ jps
4465 NodeManager
4356 DataNode
4584 Jps
slave2节点
[hadoop@slave2 ~]$ jps
3714 DataNode
3942 Jps
3823 NodeManager
步骤十三:在所有节点启动ZooKeeper
master节点
[hadoop@master ~]$ zkServer.sh start
[hadoop@master ~]$ jps
10288 ResourceManager
9939 NameNode
10599 Jps
10136 SecondaryNameNode
10571 QuorumPeerMain
slave1节点
[hadoop@slave1 ~]$ zkServer.sh start
[hadoop@slave1 ~]$ jps
1473 QuorumPeerMain
1302 NodeManager
1226 DataNode
1499 Jps
slave2节点
[hadoop@slave2 ~]$ zkServer.sh start
[hadoop@slave2 ~]$ jps
1296 NodeManager
1493 Jps
1222 DataNode
1469 QuorumPeerMain
步骤十四:在master节点启动HBase
注意:各节点之间时间必须同步,否则HBase启动不了。
在每个节点执行date命令,查看每个节点的时间是否同步,不同步的话,在各节点执行date命令,date
-s "2016-04-15 12:00:00"。
master节点
[hadoop@master ~]$ start-hbase.sh
[hadoop@master ~]$ jps
1669 ResourceManager
2327 Jps
1322 NameNode
2107 HMaster
1948 QuorumPeerMain
1517 SecondaryNameNode
slave1节点
[hadoop@slave1 ~]$ jps
1473 QuorumPeerMain
1557 HRegionServer
1702 Jps
1302 NodeManager
1226 DataNode
slave2节点
[hadoop@slave2 ~]$ jps
1296 NodeManager
1222 DataNode
1545 HRegionServer
1725 Jps
1469 QuorumPeerMain
步骤十五:在浏览器输入master:60010出现如下图7-2所示的界面。
{width="5.768055555555556in"
height="2.8243055555555556in"}
图7-2 HBase浏览器界面
注意:若slave显示不全,具体问题查看日志解决
实验任务二:HBase常用Shell命令
启动hdfs、zookeeper、hbase服务
步骤一:进入HBase命令行
[hadoop@master ~]$ hbase shell
步骤二:建立表scores,两个列簇:grade和course
hbase(main):001:0> create 'scores','grade','course'
0 row(s) in 1.4480 seconds
=> Hbase::Table - scores
步骤三:查看数据库状态
hbase (main) :001 :0> status
1 active master, 0 backup masters, 2 servers, 0 dead, 1.0000 average load
步骤四:查看数据库版本
hbase (main) :002:0> version
1.2.1,r8d8a7107dc4ccbf36a92f64675dc60392f85c015,Wed Mar 30 11:19:21 CDT 20
16
步骤五:查看表
hbase(main):008:0> list
TABLE
scores
1 row(s) in 0.0100 seconds
=>["scores"]
步骤六:插入记录1:zhangsan,grade: 143cloud
hbase(main):003:0> put 'scores','zhangsan','grade:','146cloud'
0 row(s) in 0.2250 seconds
步骤七:插入记录2:zhangsan,course:math,86
hbase(main):004:0> put 'scores','zhangsan','course:math','86'
0 row(s) in 0.0190 seconds
步骤八:插入记录3:zhangsan,course:cloud,92
hbase(main):005:0> put 'scores','zhangsan','course:cloud','92'
0 row(s) in 0.0170 seconds
步骤九:插入记录4:lisi,grade:133soft
hbase(main):006:0> put 'scores','lisi','grade:','133soft'
0 row(s) in 0.0070 seconds
步骤十:插入记录5:lisi,grade:math,87
hbase(main):007:0> put 'scores','lisi','course:math','87'
0 row(s) in 0.0060 seconds
步骤十一:插入记录6:lisi,grade:cloud,96
hbase(main):008:0> put 'scores','lisi','course:cloud','96'
0 row(s) in 0.0070 seconds
步骤十二:读取zhangsan的记录
hbase(main):009:0> get 'scores','zhangsan'
COLUMN CELL
course:c Loud timestamp= 1460479208148, value=92
course : math timestamp=1460479163325,value=86
grade : timestamp=1460479064086,value= 146cloud
3 row(s) in 0.0800 seconds
步骤十三:读取zhangsan的班级
hbase(main):012:0> get 'scores','zhangsan','grade'
COLUMN CELL
grade : timestamp= 1460479064086,value=146cloud
1 row( s) in 0.0150 seconds
步骤十四:查看整个表记录
hbase(main):013:0> scan 'scores'
ROW COLUMN+CELL
lisi column=course:cloud, timestamp=1598492107507, value=96
lisi column=course:math, timestamp=1598492101697, value=87
lisi column=grade:, timestamp=1598492094568, value=133soft
zhangsan column=course:cloud, timestamp=1598492087700, value=92
zhangsan column=course:math, timestamp=1598492077530, value=86
zhangsan column=grade:, timestamp=1598492061508, value=146cloud
2 row(s) in 0.0570 seconds
步骤十五:按例查看表记录
hbase(main):014:0> scan 'scores',{COLUMNS=>'course'}
ROW COLUMN+CELL
zhangsan column=course:cloud, timestamp= 1460479208148 ,value=92
zhangsan column=course:math, timestamp=1460479163325,value=86
lisi column=course:cloud, timestamp=1460479342925,value=96
lisi column=course:math, times tamp=1460479312963, value=87
2 row(s) in 0. 0230 seconds
步骤十六:删除指定记录
hbase(main):015:0> delete 'scores','lisi','grade'
0 row(s) in 0.0390 seconds
步骤十七:删除后,执行scan命令
hbase(main):016:0> scan 'scores'
ROW COLUMN+CELL
zhangsan column=course: cloud, timestamp=1460479208148,value=92
zhangsan column=course :math, timestamp= 1460479163325,value=86
zhangsan co lumn=grade: ,timestamp= 1460479064086,value=146cloud
lisi column=course :cloud, timestamp= 1460479342925, value=96
lisi column=course : math, timestamp= 1460479312963, value=87
row( s) in 0. 0350 seconds
步骤十八:增加新的列簇
hbase(main):017:0> alter 'scores',NAME=>'age'
Updating all regions with the new schema.
0/ 1 regions updated.
1/ 1 regions updated.
Done.
0 row(s) in 3.0060 seconds
步骤十九:查看表结构
hbase(main):018:0> describe 'scores'
Table scores is ENABLED
scores
COLUMN FAMIL IES DESCRIPTION
{NAME => age',BL O0MFILTER =>ROW',VERSIONS => '1' ,IN_ MEMORY => 'false', KEEP DELETED_ CELLS =>FAL SE',DATA BLOCK ENCODING =>
NONETTL => ' FOREVER', COMPRESSION=>NONE',MIN VERSIONS => '0' ,
BLOCKCACHE => ' true BLOCKSIZE =>65536',REPLICATION_ SCOPE => '0'}
{NAME =>course,BLOOMFILTER => ' ROW', VERSIONS =>IN MEMORY =>
false', KEEPDELETED CELLS =>FALSE', DATA BLOCK ENCODING =>
1 NONE',TTL =>FOREVER',COMPRESSION =>NONE',MIN VERSIONS => '0'
BLOCKCACHE => 'true' ,BLOCKSIZE =>65536',REPLICATION SCOPE =>0'}{NAME => grade BLOOMFILTER => 'ROW' ,VERSIONS =>IN_ MEMORY => ' false', KEEPDELETED CELLS =>FALSE',DATA BLOCK ENCODING =>1 NONE 'TTL =>FOREVER',COMPRESSION =>NONE',MIN VERSIONS => '0'
BLOCKCACHE => 'true' ,BLOCKSIZE => '65536 ,REPLICATION SCOPE => '0'}
3 row(s) in 0.0400 seconds
步骤二十:删除列簇
hbase(main):020:0> alter 'scores',NAME=>'age',METHOD=>'delete'
Updating all regions with the new schema…
1/ 1 regions updated.
Done,
0 row(s)in 2.1600seconds
步骤二十一:删除表
hbase(main):021:0> disable 'scores'
0 row(s)in 2.2930seconds
hbase(main):022:0> drop 'scores'
0 row(s)in 1.2530seconds
hbase( main) :023:0> list
TABLE
0 row(s)in 0.0150 seconds
==>[]
步骤二十二:退出
hbase(main):024:0> quit
[hadoopdmaste r~]$
步骤二十三:关闭HBase
在master节点关闭HBase。
[hadoop@master ~]$ stop-hbase.sh
在所有节点关闭ZooKeeper。
[hadoop@master ~]$ zkServer.sh stop
[hadoop@slave1 ~]$ zkServer.sh stop
[hadoop@slave2 ~]$ zkServer.sh stop
在master节点关闭Hadoop。
[hadoop@master ~]$ stop-all.sh