集群环境下配置hadoop1.0,zookeeper,hbase

1.本环境采用两台linux环境，ip分别为： 
192.168.56.101 
192.168.56.102 
分别修改两台机器的/etc/hosts文件，增加如下内容： 
192.168.56.101 master 
192.168.56.102 slave 
我们以master作为namenode服务器，slave为datenode服务器，首先安装jdk并配置环境变量和ssh（安装过程省略，请参考 

网上资料） 
2.安装hadoop,修改如下配置文件： 
两台机器的hadoop安装路径要相同，切记，切忌！！！ 
1）修改core-site.xml为： 
<?xml version="1.0"?> 
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?> 

<!-- Put site-specific property overrides in this file. --> 

<configuration> 
    <property> 
        <name>fs.default.name</name> 
        <value>hdfs://master:9000</value> 
    </property> 
</configuration> 
2）修改hdfs-site.xml为： 
<?xml version="1.0"?> 
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?> 

<!-- Put site-specific property overrides in this file. --> 

<configuration> 
    <property> 
        <name>dfs.replication</name> 
        <value>1</value> 
    </property> 
    <property> 
        <name>dfs.support.append</name> 
        <value>true</value> 
    </property> 
    <property> 
        <name>dfs.datanode.max.xcievers</name> 
        <value>4096</value> 
    </property> 
</configuration> 
3）修改mapred-site.xml为： 
<?xml version="1.0"?> 
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?> 

<!-- Put site-specific property overrides in this file. --> 

<configuration> 
    <property> 
        <name>mapred.job.tracker</name> 
        <value>master:9001</value> 
    </property> 
</configuration> 
4)修改masters为： 
master 
5）修改slaves为： 
slave 
6)格式化hadoop文件系统： 
hadoop namenode -format 
7)启动hadoop： 

start-all.sh

3.安装zookeeper,修改配置文件： 
两台机器的zookeeper安装路径要相同，切记，切忌！！！ 
1）cp zoo_sample.cfg zoo.cfg 
2)修改zoo.cfg为： 
# 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. 
# do not use /tmp for storage, /tmp here is just 
# example sakes. 
dataDir=/tmp/zookeeper 
# the port at which the clients will connect 
clientPort=2181 
#... 
#... 
#... 
server.1=master:2888:3888 
server.2=slave:2888:3888 
3）在master的/tmp/zookeeper路径下增加myid文件，内容为：1 
4）在master的/tmp/zookeeper路径下增加myid文件，内容为：2 
补充说明：如果要单独使用zookeeper集群服务器，则使用zkServer.sh start命令分别启动每个zookeeper节点 

4.安装hbase，修改配置文件 
1）修改：hbase_env.sh的两个属性： 
# The java implementation to use.  Java 1.6 required. 
export JAVA_HOME=/opt/java 
# Tell HBase whether it should manage it's own instance of Zookeeper or not. 
export HBASE_MANAGES_ZK=true 
2）在两台机器上的hadoop文件系统中创建hbase目录： 
hadoop fs -mkdir /hbase 
3）修改master上的hbase-site.xml文件： 
<configuration> 
    <property> 
        <name>dfs.support.append</name> 
        <value>true</value> 
    </property> 
    <property> 
    <name>hbase.rootdir</name> 
    <value>hdfs://master:9000/hbase</value> 
    <description>The directory shared by RegionServers. 
    </description> 
  </property> 
  <property> 
    <name>dfs.replication</name> 
    <value>1</value> 
    <description>The replication count for HLog and HFile storage. Should not be greater than HDFS datanode count. 
    </description> 
  </property> 
  <property>  
      <name>hbase.cluster.distributed</name> 
      <value>true</value>  
   </property>  
   <property>  
           <name>hbase.master.port</name>  
           <value>60000</value>  
   </property>  
   <property>  
           <name>hbase.zookeeper.quorum</name>  
           <value>master,slave</value>  
   </property> 
</configuration> 
4）修改slave上的hbase-site.xml文件： 
<configuration> 
    <property> 
        <name>dfs.support.append</name> 
        <value>true</value> 
    </property> 
    <property> 
    <name>hbase.rootdir</name> 
    <value>hdfs://master:9000/hbase</value> 
    <description>The directory shared by RegionServers. 
    </description> 
  </property> 
  <property> 
    <name>dfs.replication</name> 
    <value>1</value> 
    <description>The replication count for HLog and HFile storage. Should not be greater than HDFS datanode count. 
    </description> 
  </property> 
  <property>  
      <name>hbase.cluster.distributed</name> 
      <value>true</value>  
   </property>  
   <property>  
           <name>hbase.master.port</name>  
           <value>60000</value>  
   </property>  
   <property>  
           <name>hbase.zookeeper.quorum</name>  
           <value>master,slave</value>  
   </property>  
</configuration> 
5)修改regionservers文件为： 
master 
slave 
6)启动hbase并测试： 
start-hbase.sh 
hbase shell 
list 
哈哈，完毕！