hbase集群安装（hadoop0.20.2+zookeeper3.3.3+hbase0.90.2）

并且已经安装成功Hadoop-0.20.2及以上版本。

安装包准备

需要安装包：

zookeeper-3.3.2.tar.gz（stable版本）

hbase-0.20.2.tar.gz（stable版本）

安装步骤

安装和配置ZooKeeper

HBase从0.20.0开始，需要首先安装ZooKeeper。从apache上下载zookeeper-3.2.1.tar.gz（Stable版本），解压到/home/hdfs/目录下。

（1）在namenode节点新建zookeeper目录，在该目录下新建myid文件。

（2）在zookeeper-3.2.1/conf目录下，拷贝zoo_sample.cfg为zoo.cfg。在zoo.cfg中将dataDir改为/home/hdfs/zookeeper，在文件末位添加所有的主机：

server.1=10.192.1.1:2888:3888

server.2=10.192.1.2:2888:3888

server.3=10.192.1.3:2888:3888

 

（3）用scp命令将namenode节点的的/home/hdfs/ zookeeper-3.2.1和/home/hdfs/ zookeeper拷贝到其余所有主机的/home/hdfs目录下。

（4）参照zoo.cfg中的配置，在各主机myid文件中写入各自的编号。如：10.192.1.1入1,10.192.1.2写入2

（5）在所有节点上执行bin/zkServer.sh start，分别启动。

执行bin/zkCli.sh -server xxx.xxx.xxx.xxx:2181，检查指定服务器是否成功启动。

vi /etc/profile
export HBASE_HOME=/hadoop/hbase
export PATH=$PATH:$HBASE_HOME/bin

export HADOOP_HOME=/hadoop/hadoop
export PATH=$PATH:$HADOOP_HOME/bin

 

安装和配置HBase

tar -zxvf hbase-0.20.6.tar.gz

cd hbase-0.20.6

mv * /hadoop/hbase

下载HBase0.20.1版本，解压到namenode节点的/home/hdfs目录下。

配置说明

（1）系统所有配置项的默认设置在hbase-default.xml中查看，如果需要修改配置项的值，在hbase-site.xml中添加配置项。

在分布式模式下安装HBase，需要添加的最基本的配置项如下：

<property>

<name>hbase.rootdir</name>

<value>hdfs://hadoop2-namenode:9000/hbase</value>

<description>The directory shared by region servers.</description>

</property>

<property>

<name>hbase.cluster.distributed</name>

<value>true</value>

<description>The mode the cluster will be in. Possible values are

false: standalone and pseudo-distributed setups with managed Zookeeper

true: fully-distributed with unmanaged Zookeeper Quorum (see hbase-env.sh)

</description>

</property>
<property>
      <name>hbase.zookeeper.property.dataDir</name>
      <value>/hadoop/zookeeper</value>
      <description>Property from ZooKeeper's config zoo.cfg.
      The directory where the snapshot is stored.
      </description>
    </property>
 <property>
      <name>hbase.zookeeper.property.clientPort</name>
      <value>2222</value>
      <description>Property from ZooKeeper's config zoo.cfg.
      The port at which the clients will connect.
      </description>
    </property>

<property>
      <name>hbase.zookeeper.quorum</name>
      <value>hadoop2-namenode,hadoop2-datanode1,hadoop2-datanode2</value>
      <description>Comma separated list of servers in the ZooKeeper Quorum.
      For example, "host1.mydomain.com,host2.mydomain.com,host3.mydomain.com".
      By default this is set to localhost for local and pseudo-distributed modes
      of operation. For a fully-distributed setup, this should be set to a full
      list of ZooKeeper quorum servers. If HBASE_MANAGES_ZK is set in hbase-env.sh
      this is the list of servers which we will start/stop ZooKeeper on.
      </description>
    </property>

 

（2）在conf/hbase-env.sh中修改添加配置项：
export JAVA_HOME=/usr/java/jdk1.6.0_22

 

 export HBASE_MANAGES_ZK=true

 
export HADOOP_CONF_DIR=/hadoop/hadoop-config

HADOOP_CLASSPATH 要設hbase 的環境，而HBASE_CLASSPATH要設hadoop的環境，有了這行可以解決編譯hbase 程式時出現run time error

并把~/hadoop-0.20.1/conf/hdfs-site.xml拷贝至~/hbase-3.2.1/conf/目录下。

（3）将ZooKeeper的配置文件zoo.cfg添加到HBase所有主机的CLASSPATH中。

（4）在conf/regionservers中添加hadoop-0.20.1/conf/slaves中所有的datanode节点。
hadoop2-datanode1
hadoop2-datanode2

启动

Hadoop、ZooKeeper和HBase之间应该按照顺序启动和关闭：启动Hadoop—>启动ZooKeeper集群—>启动HBase—>停止HBase—>停止ZooKeeper集

群—>停止Hadoop。

在namenode节点执行bin/hbase-daemon.sh，启动master。执行bin/start-hbase.sh和bin/stop-hbase.sh 脚本启动和停止HBase服务。

/hadoop/hbase/bin/hbase-daemon.sh start master
/hadoop/hbase/bin/hbase-daemon.sh stop master
/hadoop/hbase/bin/start-hbase.sh
/hadoop/hbase/bin/stop-hbase.sh

/hadoop/hbase/bin/hbase shell

 

 

================================================

环境准备

 1.在windows下安装VMware
 
 2.创建了3个fedora14 linux。地址分别为：
  m201 192.168.0.201 (Namenode)
  s202 192.168.0.202 (Datanode)
  s203 192.168.0.203 (Datanode)
 
 3.在linux系统中下载所需要的软件。分别为：
  jdk-6u23-linux-i586-rpm.bin
  hadoop-0.20.2.tar.gz
  zookeeper-3.3.3.tar.gz
  hbase-0.90.2.tar.gz
 将下载的软件保存到/root/install目录下。

 

安装jdk(s202,s203进行同样的操作)

 1.执行jdk-6u23-linux-i586-rpm.bin就行可以。jdk将安装在/usr/java/jdk1.6.0_23目录下。
 
 2.设置java环境变量，修改/etc/profile文件。在文件最后增加：
  export JAVA_HOME=/usr/java/jdk1.6.0_23/
  export JRE_HOME=/usr/java/jdk1.6.0_23/jre/
  export CLASSPATH=.:$JAVA_HOME/lib:$JRE_HOME/lib:$CLASSPATH
  export PATH=$JAVA_HOME/bin:$PATH
 3.使/etc/profile文件生效，执行这个文件。

 

设置ssh(使m201,可以不用密码访问s202和s203)
 
 官网上的一段话：

 Now check that you can ssh to the localhost without a passphrase:
 $ ssh localhost

 If you cannot ssh to localhost without a passphrase, execute the following commands:
 $ ssh-keygen -t dsa -P '' -f ~/.ssh/id_dsa
 $ cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys

 将m201上的id_dsa.pub 文件追加到s202和s203的authorized_keys文件内

安装hadoop
 
 1.到/root/install目录解压hadoop-0.20.2.tar.gz，执行命令：tar -zxvf hadoop-0.20.2.tar.gz。运行结束后将生成hadoop-0.20.2目录

 2。进入/root/install/hadoop-0.20.2/conf目录

 3.修改文件masters(定义masters IP)
  192.168.0.201
 
 4.修改文件slaves(定义slaves IP)
  192.168.0.202
  192.168.0.203
  
 5.修改文件hadoop-env.sh(设置jdk路径)
  export JAVA_HOME=/usr/java/jdk1.6.0_23
  
 6.修改文件core-site.xml在<configuration>中加入
  <property>
    <name>hadoop.tmp.dir</name>
    <value>/hadoopdata</value>
    <description>A base for other temporary directories.</description>
  </property>

  <property>
    <name>fs.default.name</name>
    <value>hdfs://m201:9000</value>
    <description>The name of the default file system.  A URI whose
    scheme and authority determine the FileSystem implementation.  The
    uri's scheme determines the config property (fs.SCHEME.impl) naming
    the FileSystem implementation class.  The uri's authority is used to
    determine the host, port, etc. for a filesystem.</description>
  </property>
 
 7.修改文件hdfs-site.xml在<configuration>中加入
  <property>
   <name>dfs.replication</name>
   <value>1</value>
  </property>
 
 8.修改文件mapred-site.xml在<configuration>中加入
  <property>
    <name>mapred.job.tracker</name>
    <value>m201:9001</value>
    <description>The host and port that the MapReduce job tracker runs
    at.  If "local", then jobs are run in-process as a single map
    and reduce task.
    </description>
  </property>
 
 9.设置环境变量,修改文件/etc/profile
  export HADOOP_HOME=/root/install/hadoop-0.20.2
  export PATH=$HADOOP_HOME/bin:$PATH

 s202,s203,也执行一样的操作

 执行/etc/profile使其生效

 10.配/etc/hosts文件，加入
  192.168.0.201 m201
  192.168.0.202 s202
  192.168.0.203 s203
 s202,s203,也执行一样的操作
 
 11.将/root/install/hadoop-0.20.2目录复制到s202,s203上
  可使用scp -r 源 主机:目标

 11.格式化HDFS文件系统
   /root/install/hadoop-0.20.2/bin/hadoop namenode –format命令

 12. 执行/root/install/hadoop-0.20.2/bin/start-all.sh文件，启服务
  /root/install/hadoop-0.20.2/bin/stop-all.sh文件，停止服务
  
 hadoop安装完成
  可运行
  http://192.168.0.201:50070/dfshealth.jsp
  查看hadoop是否运行

安装zookeeper(在m201上执行)

 1.在/root/install/hadoop-0.20.2/中创建目录zookeeper
  cd /root/install/hadoop-0.20.2
  mkdir zookeeper
 
 2.在/root/install目录中解压zookeeper
  cd /root/install
  tar -zxvf zookeeper-3.3.3.tar.gz

 3.将zookeeper移动至/root/install/hadoop-0.20.2/zookeeper目录
  cd /root/install/zookeeper
  mv * /root/install/hadoop-0.20.2/zookeeper
 
 3配置zookeeper
  1).创建zoo.cfg文件
   cd /root/install/hadoop-0.20.2/zookeeper/conf
   cp zoo_sample.cfg zoo.cfg
  2).修改zoo.cfg文件，zoo.cfg文件的完整内容如下：
   # The number of milliseconds of each tick
   tickTime=2000
   # The number of ticks that the initial
   # synchronization phase can take
   initLimit=10
   # The number of ticks that can pass between
   # sending a request and getting an acknowledgement
   syncLimit=5
   # the directory where the snapshot is stored.

   dataDir=/root/install/hadoop-0.20.2/zookeeper/zookeeper-data #(新增加)

   dataLogDir=/root/install/hadoop-0.20.2/zookeeper/logs  #(新增加)

   # the port at which the clients will connect
   clientPort=2181

   server.1=m201:2888:3888      #(新增加)

   server.2=s202:2888:3888      #(新增加)

   server.3=s203:2888:3888      #(新增加)

   在文件中写入 #(新增加)的项目

   
  
  3).创建zookeeper-data目录
   cd /root/install/hadoop-0.20.2/zookeeper/
   mkdir zookeeper-data

  3).创建myid文件
   cd /root/install/hadoop-0.20.2/zookeeper/zookeeper-data
   vi myid
    myid文件中的内空写:1
   :x保存文件

 4.将/root/install/hadoop-0.20.2/zookeeper目录复制到s202,s203上
  可使用scp -r 源 主机:目标

 5.进入s202主机，写myid文件内容修改为:2

 6.进入s203主机，写myid文件内容修改为:3

 7.启动zookeeper(m201,s202,s203，执行同样的操作)
  /root/install/hadoop-0.20.2/zookeeper/bin/zkServer.sh start
  /root/install/hadoop-0.20.2/zookeeper/bin/zkServer.sh stop(为停止)

 

安装hbase(m201中操作)
 1.在/root/install/hadoop-0.20.2/中创建目录hbase
  cd /root/install/hadoop-0.20.2
  mkdir hbase
 
 2.在/root/install目录中解压hbase
  cd /root/install
  tar -zxvf hbase-0.90.2.tar.gz

 3.将hbase移动至/root/install/hadoop-0.20.2/hbase目录
  cd /root/install/hbase-0.90.2
  mv * /root/install/hadoop-0.20.2/hbase

 4.配置hbase
  1).配置/etc/profile文件，加入
   export HBASE_HOME=/root/install/hadoop-0.20.2/hbase
   export PATH=$PATH:$HBASE_HOME/bin

  s202,s203,也执行一样的操作

  执行/etc/profile使其生效

  2).修改hbase-site.xml文件
   cd /root/install/hadoop-0.20.2/hbase/conf
   vi hbase-site.xml

  在<configuration>中加入 :
   <property>

    <name>hbase.rootdir</name>

    <value>hdfs://m201:9000/hasexx</value>

    <description>The directory shared by region servers.</description>

   </property>

     <property>
       <name>hbase.master.port</name>
       <value>60000</value>
     </property>

   <property>

    <name>hbase.cluster.distributed</name>

    <value>true</value>

    <description>The mode the cluster will be in. Possible values are

    false: standalone and pseudo-distributed setups with managed Zookeeper

    true: fully-distributed with unmanaged Zookeeper Quorum (see hbase-env.sh)

    </description>

   </property>

   <property>

       <name>dfs.replication</name>

       <value>1</value>

    </property>

   <property>
         <name>hbase.zookeeper.property.dataDir</name>
         <value>/root/install/hadoop-0.20.2/zookeeper</value>
         <description>Property from ZooKeeper's config zoo.cfg.
         The directory where the snapshot is stored.
         </description>
       </property>
    <property>
         <name>hbase.zookeeper.property.clientPort</name>
         <value>2181</value>
         <description>Property from ZooKeeper's config zoo.cfg.
         The port at which the clients will connect.
         </description>
       </property>

   <property>
         <name>hbase.zookeeper.quorum</name>
         <value>m201,s202,s203</value>
         <description>Comma separated list of servers in the ZooKeeper Quorum.
         For example, "host1.mydomain.com,host2.mydomain.com,host3.mydomain.com".
         By default this is set to localhost for local and pseudo-distributed modes
         of operation. For a fully-distributed setup, this should be set to a full
         list of ZooKeeper quorum servers. If HBASE_MANAGES_ZK is set in hbase-env.sh
         this is the list of servers which we will start/stop ZooKeeper on.
         </description>
    </property>

 

  3).修改hbase-env.sh文件,加入
   export JAVA_HOME=/usr/java/jdk1.6.0_23/
   export HBASE_CLASSPATH=/root/install/hadoop-0.20.2/conf
   export HBASE_MANAGES_ZK=false

  4).复制zookeeper的zoo.cfg文件到/root/install/hadoop-0.20.2/conf目录中
   cp /root/install/hadoop-0.20.2/zookeeper/conf/zoo.cfg /root/install/hadoop-0.20.2/conf/
   

  5).修改regionservers文件，完整内容为：
   192.168.0.202
   192.168.0.203

  6).将hadoop的hadoop-0.20.2-core.jar文复制到hbase的lib目录下，删除原来的hadoop-core-0.20-append-r1056497.jar文件

  7).将/root/install/hadoop-0.20.2/hbase目录复制到s202,s203上
   可使用scp -r 源主机:目标

 5.启动服务
  /root/install/hadoop-0.20.2/hbase/bin/start-hbase.sh
  /root/install/hadoop-0.20.2/hbase/bin/stop-hbase.sh停止

  
 
 http://192.168.0.201:60010/master.jsp
 http://192.168.0.202:60030/regionserver.jsp
 http://192.168.0.203:60030/regionserver.jsp