软件版本号:
JDK:jdk-8u45-linux-i586.tar.gz
Zookeeper:zookeeper-3.4.6
Hbase:hbase-1.0.0-bin
一、JDK版本更换
由于之前安装的jdk-6u45-linux-i586版本较低,与Hbase不兼容,因此更换为jdk-8u45-linux-i586
1.jdk 1.6卸载
(1)删除jdk目录:sudo rm -rf /opt/jdk1.6.0_45
(2)删除jdk环境变量:sudo gedit /etc/profile
删除代码段:
# java
export JAVA_HOME=/opt/jdk1.6.0_45
export JRE_HOME=$JAVA_HOME/jre
export CLASSPATH=$JAVA_HOME/lib:$JRE_HOME/lib:$CLASSPATH
export PATH=$JAVA_HOME/bin:$JRE_HOME/bin:$PATH
2.jdk 1.8安装
(1)复制jdk-8u45-linux-i586.tar.gz到/opt目录下并解压为文件夹jdk
(2)配置jdk环境变量:sudo gedit /etc/profile
添加代码段:
# java
export JAVA_HOME=/opt/jdk1.8.0_45
export JRE_HOME=$JAVA_HOME/jre
export CLASSPATH=$JAVA_HOME/lib:$JRE_HOME/lib:$CLASSPATH
export PATH=$JAVA_HOME/bin:$JRE_HOME/bin:$PATH
(3)执行命令,使配置生效:source /etc/profile;
(4)执行命令:java -version,若出现java版本号,则说明安装成功。
(5)修改/home/hduser/hadoop/etc/hadoop路径下hadoop-env.sh和yarn-env.sh中jdk的版本号
(6)Ubuntu2执行相同操作
二、Zookeeper配置
1.Zookeeper配置
(1)复制zookeeper-3.4.6.tar.gz到/home/hduser目录下并解压为文件夹zookeeper-3.4.6
(2)配置Zookeeper环境变量:sudo gedit /etc/profile
添加代码段:
#zookeeper
export ZOOKEEPER_HOME=/home/hduser/zookeeper-3.4.6
export PATH=$ZOOKEEPER_HOME/bin:$ZOOKEEPER_HOME/conf:$PATH
(3)新建zoo.cfg并修改:
hduser@Ubuntu1:~/zookeeper-3.4.6$ cp conf/zoo_sample.cfg conf/zoo.cfg
清空zoo.cfg中代码,然后添加以下代码段:
tickTime=2000
initLimit=10
syncLimit=5
dataDir=/home/hduser/zookeeper-3.4.6/zookeeper
dataLogDir=/home/hduser/zookeeper-3.4.6/log
clientPort=2181
server.1=Ubuntu1:2888:3888
server.2=Ubuntu2:2888:3888
(4)在/home/hduser/zookeeper-3.4.6目录下新建文件夹zookeeper,在其中新建文件myid,并写入“1”
(5)Ubuntu2执行相同操作,不同点是myid写入的值为“2”。
2.Zookeeper测试
(1)各节点启动:
hduser@Ubuntu1:~/zookeeper-3.4.6$ bin/zkServer.sh start
hduser@Ubuntu2:~/zookeeper-3.4.6$ bin/zkServer.sh start
(2)jps查看:3347 QuorumPeerMain
(3)查看状态:
hduser@Ubuntu1:~/zookeeper-3.4.6$ bin/zkServer.sh status
JMX enabled by default
Using config: /home/hduser/zookeeper-3.4.6/bin/../conf/zoo.cfg
Mode: follower
hduser@Ubuntu2:~/zookeeper-3.4.6$ bin/zkServer.sh status
JMX enabled by default
Using config: /home/hduser/zookeeper-3.4.6/bin/../conf/zoo.cfg
Mode: leader
(4)停止ZooKeeper:
hduser@Ubuntu1:~/zookeeper-3.4.6$ bin/zkServer.sh stop
三、Hbase配置
1.Hbase分布式配置
(1)下载安装包hbase-1.0.0-bin.tar.gz并复制到/home/hduser目录下并解压为文件夹hbase-1.0.0
(2)修改conf/hbase-site.xml文件:
添加代码段:
hbase.rootdir
hdfs://Ubuntu1:9000/hbase
hbase.cluster.distributed
true
hbase.zookeeper.property.clientPort
2181
hbase.zookeeper.quorum
Ubuntu1,Ubuntu2
hbase.zookeeper.property.dataDir
/home/hduser/zookeeper-3.4.6/zookeeper
dfs.support.append
true
(3)修改conf/regionservers文件:
添加:Ubuntu1
Ubuntu2
(4)修改conf/hbase-env.sh文件:
添加:
export JAVA_HOME=/opt/jdk1.7.0_07
export HBASE_MANAGES_ZK=false
export HBASE_HOME=/home/hduser/hbase-1.0.0
export HADOOP_HOME=/home/hduser/hadoop
(5)修改各节点Hadoop的hdfs-site.xml文件:
添加:
dfs.datanode.max.xcievers
4096
注:该参数限制了datanode所允许同时执行的发送和接受任务的数量,缺省为256
(6)Ubuntu2执行相同操作
Hbase配置
1.Hbase配置单机
(1)下载安装包hbase-1.0.0-bin.tar.gz并复制到到/home/hduser目录下并解压为文件夹hbase-1.0.0
(2)修改conf/hbase-site.xml文件:
添加代码段:
hbase.rootdir
hdfs://Ubuntu1:9000/hbase
hbase.cluster.distributed
false
hbase.zookeeper.property.clientPort
2181
hbase.zookeeper.quorum
Ubuntu1
hbase.zookeeper.property.dataDir
/home/hduser/zookeeper-3.4.6/zookeeper
dfs.support.append
true
(3)修改conf/regionservers文件:
添加:Ubuntu1
(4)修改conf/hbase-env.sh文件:
添加:
export JAVA_HOME=/opt/jdk1.8.0_45
export HBASE_MANAGES_ZK=true
export HBASE_HOME=/home/hduser/hbase-1.0.0
export HADOOP_HOME=/home/hduser/hadoop
(5)修改各节点Hadoop的hdfs-site.xml文件:
添加:
dfs.datanode.max.xcievers
4096
注:该参数限制了datanode所允许同时执行的发送和接受任务的数量,缺省为256
2.Hbase测试
启动顺序:Hadoop→Zookeeper→Hbase,结束顺序:Hbase→Zookeeper→Hadoop
(1)启动Hbase:
hduser@Ubuntu1:~/hbase-1.0.0$ bin/start-hbase.sh
(2)jps查看:
# 主节点 Ubuntu1: jps
3491 HMaster
2887 NameNode
3640 HRegionServer
3347 QuorumPeerMain
3010 DataNode
3727 Jps
3168 SecondaryNameNode
# 从节点Ubuntu2:
2870 DataNode
3280 Jps
2953 QuorumPeerMain
3054 HRegionServer
(3)进入hbase shell:
hduser@Ubuntu1:~/hbase-1.0.0$ bin/hbase shell
(4)查看集群状态
hbase(main):001:0> status
(5)建表测试:
hbase(main):003:0> create 'test1','id1'
0 row(s) in 1.8420 seconds
=>Hbase::Table - test1
=>Hbase::Table - test1
hbase(main):004:0> list
TABLE
test
test1
2 row(s) in 0.1770 seconds
=> ["test", "test1"]
注:如果jps之后显示没有安装jdk,则执行source /etc/profile后再jps
其他
注意:有时会出现Can't get master address from ZooKeeper; znode data == null
解决方法是关掉Hbase 关掉Zookeeper 删除zookeeper数据目录下相关文件(myid文件保留),启动zookeeper ,启动Hbase,
如果不行,还可以关掉hadoop,格式化namenode。
还不行,查看hbase的启动日志。可能是端口被占用,用相关命令查询。比如下面日志例子:
Could not start ZK at requested port of 2181. ZK was started at port:2182. Aborting as clients(e.g. shell) will not be able to find this ZK quorum.
说明进程占用了默认的2181端口导致ZK不能正常启动。使用lsof -i:2181命令查看2181端口的进程情况:
如果Hadoop用户的java进程在使用。Kill对应进程,,重新启动Hbase shell。
启动HDFS:./sbin/start-dfs.sh
停止HDFS:./sbin/stop-dfs.sh
启动yarn:./sbin/start-yarn.sh
停止yarn:./sbin/stop-yarn.sh
启动Zookeeper:bin/zkServer.sh start
停止Zookeeper:bin/zkServer.sh stop
启动Hbase:bin/start-hbase.sh
停止Hbase:bin/stop-hbase.sh