ubuntu11.04上cloudera cdh3u0的hadoop和hbase分布式安装

概述：

业务需要较实时的处理大数据量，并提供大吞吐量的读写，hbase作为一个可选的列数据库，记录一下安装过程，待查。 
hbase的安装需要hadoop和zookeeper和hbase，生产环境下需要将zookeeper独立安装，并保证整个集群没有单点。 
hbase软件选择cloudera的cdh3u0，apache的版本需要重新编译hadoop0.20.2-appender版以保证没有数据丢失。 

1、规划机器 
使用12个虚拟机作为集群测试，ip为192.168.0.221-232，hostname为ubuntu-1到ubuntu-12其中 

机器名	IP	安装系统
ubuntu-1	192.168.0.221	hadoop namenode / hbase HMaster
ubuntu-2	192.168.0.222	hadoop datanode / hbase HRegionServer
ubuntu-3	192.168.0.223	hadoop datanode / hbase HRegionServer
ubuntu-4	192.168.0.224	hadoop datanode / hbase HRegionServer
ubuntu-5	192.168.0.225	hadoop datanode / hbase HRegionServer
ubuntu-6	192.168.0.226	hadoop datanode / hbase HRegionServer
ubuntu-7	192.168.0.227	hadoop datanode / hbase HRegionServer
ubuntu-8	192.168.0.228	hadoop datanode / hbase HRegionServer
ubuntu-9	192.168.0.229	zookeeper
ubuntu-10	192.168.0.230	zookeeper
ubuntu-11	192.168.0.231	zookeeper
ubuntu-12	192.168.0.232	hadoop second namenode / hbase HMaster

2、安装操作系统及必要软件 
安装ubuntu11.04 server X64版,hadoop推荐2.6.30以上内核 
# sudo apt-get install ssh 
# sudo apt-get install vim 
# sudo apt-get install rsync 
创建hadoop用户 
在所有机器上，密码为hadoop，并加入到sudo组 
# sudo adduser hadoop 
# sudo usermod -G sudo hadoop 
安装jdk 
从oracle下载jdk-6u24-linux-x64.bin，在所有的机器上安装 
# ./jdk-6u24-linux-x64.bin 
# sudo mv jdk1.6.0_24 /usr/local 
解压cdh3u0软件 
在所有机器创建cdh3目录 
# mkdir /home/hadoop/cdh3 
在221上解压hadoop和hbase 
# tar zxvf hadoop-0.20.2-cdh3u0.tar.gz -C /home/hadoop/cdh3 
# tar zxvf hbase-0.90.1-cdh3u0.tar.gz -C /home/hadoop/cdh3 
在229上解压zookeeper 
# tar zxvf zookeeper-3.3.3-cdh3u0.tar.gz -C /home/hadoop/cdh3 
在221-228,232上修改/etc/profile 
# sudo vim /etc/profile 
添加 

 JAVA_HOME=/usr/local/jdk1.6.0_24
 JRE_HOME=$JAVA_HOME/jre
 CLASSPATH=.:$JAVA_HOME/lib:$JRE_HOME/lib:$CLASSPATH
 HADOOP_HOME=/home/hadoop/cdh3/hadoop-0.20.2-cdh3u0
 HBASE_HOME=/home/hadoop/cdh3/hbase-0.90.1-cdh3u0
 PATH=$JAVA_HOME/bin:$HADOOP_HOME/bin:$HBASE_HOME/bin:$PATH

 export JAVA_HOME JRE_HOME CLASSPATH HADOOP_HOME HBASE_HOME PATH

在229-231上修改/etc/profile 
添加 

 JAVA_HOME=/usr/local/jdk1.6.0_24
 ZOOKEEPER_HOME=/home/hadoop/cdh3/zookeeper-3.3.3-cdh3u0
 PATH=$JAVA_HOME/bin:$ZOOKEEPER_HOME/bin:$ZOOKEEPER_HOME/conf:$PATH

 export JAVA_HOME ZOOKEEPER_HOME PATH

ssh免密码登录 
用hadoop用户登录所有机器，在/home/hadoop/下建立.ssh目录 
运行 
# ssh-keygen -t dsa -P '' -f ~/.ssh/id_dsa 
这样会在~/.ssh/生成两个文件：id_dsa 和id_dsa.pub。 
# cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys 
拷贝authorized_keys到222-232 
# scp /home/hadoop/.ssh/authorized_keys [email protected]:/home/hadoop/.ssh/ 
修改hosts文件 
# sudo vim /etc/hosts 
修改所有机器的/etc/hosts文件为 

 127.0.0.1       localhost
 192.168.0.221   ubuntu-1
 192.168.0.222   ubuntu-2
 192.168.0.223   ubuntu-3
 192.168.0.224   ubuntu-4
 192.168.0.225   ubuntu-5
 192.168.0.226   ubuntu-6
 192.168.0.227   ubuntu-7
 192.168.0.228   ubuntu-8
 192.168.0.229   ubuntu-9
 192.168.0.230   ubuntu-10
 192.168.0.231   ubuntu-11
 192.168.0.232   ubuntu-12

从221登录221-232,用hostname，第一次需要确认，以后就可以直接登录。 
# ssh ubuntu-1 
# ssh ubuntu-2 
# ssh ubuntu-3 
# ssh ubuntu-4 
# ssh ubuntu-5 
# ssh ubuntu-6 
# ssh ubuntu-7 
# ssh ubuntu-8 
# ssh ubuntu-9 
# ssh ubuntu-10 
# ssh ubuntu-11 
# ssh ubuntu-12 

3、安装hadoop 
在221和232上创建/data 
# sudo mkdir /data 
# sudo chown hadoop /data 
在222和228上创建/disk1,/disk2,/disk3 
# sudo mkdir /disk1 
# sudo mkdir /disk2 
# sudo mkdir /disk3 
# sudo chown hadoop /disk1 
# sudo chown hadoop /disk2 
# sudo chown hadoop /disk3 

修改/home/hadoop/cdh3/hadoop-0.20.2-cdh3u0/conf/hadoop-env.sh添加 

 export JAVA_HOME=/usr/local/jdk1.6.0_24

修改/home/hadoop/cdh3/hadoop-0.20.2-cdh3u0/conf/core-site.xml添加 

 <property>
   <name>hadoop.tmp.dir</name>
   <value>/data</value>
   <description>A base for other temporary directories.</description>
 </property>

 <property>
   <name>fs.default.name</name>
   <value>hdfs://ubuntu-1:9000/</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>

修改/home/hadoop/cdh3/hadoop-0.20.2-cdh3u0/conf/hdfs-site.xml添加 

 <property>
   <name>dfs.name.dir</name>
   <value>/home/hadoop/data</value>
 </property>

 <property>
   <name>dfs.data.dir</name>
   <value>/disk1,/disk2,/disk3</value>
 </property>

 <property>
   <name>dfs.permissions</name>
   <value>false</value>
 </property>

 <property>
   <name>dfs.replication</name>
   <value>3</value>
 </property>

修改/home/hadoop/cdh3/hadoop-0.20.2-cdh3u0/conf/mapred-site.xml添加 

 <property>
   <name>mapred.job.tracker</name>
   <value>ubuntu-1:9001</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>

修改/home/hadoop/cdh3/hadoop-0.20.2-cdh3u0/conf/masters添加 

 ubuntu-12

修改/home/hadoop/cdh3/hadoop-0.20.2-cdh3u0/conf/slaves添加 

 ubuntu-2
 ubuntu-3
 ubuntu-4
 ubuntu-5
 ubuntu-6
 ubuntu-7
 ubuntu-8

拷贝221的hadoop到222-228,232 
# scp -r /home/hadoop/cdh3/hadoop-0.20.2-cdh3u0/ [email protected]:/home/hadoop/cdh3/ 
格式化hadoop文件系统 
# hadoop namenode -format 
启动hadoop，在221上运行 
# start-all.sh 

查看集群状态: http://192.168.0.221:50070/dfshealth.jsp 
 

查看JOB状态: http://192.168.0.221:50030/jobtracker.jsp 


4、安装zookeeper 
在229-231上创建/home/hadoop/zookeeperdata目录 
修改229的/home/hadoop/cdh3/zookeeper-3.3.3-cdh3u0/conf/zoo.cfg 

 # 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=/home/hadoop/zookeeperdata
 # the port at which the clients will connect
 clientPort=2181
 server.1=ubuntu-9:2888:3888
 server.2=ubuntu-10:2888:3888
 server.3=ubuntu-11:2888:3888

拷贝229的hadoop到230,231 
# scp -r /home/hadoop/cdh3/zookeeper-3.3.3-cdh3u0/ [email protected]:/home/hadoop/cdh3/ 
在229,230,231的/home/hadoop/zookeeperdata目录下建myid文件，内容分别为1,2,3 
启动zookeeper，在229-231上分别执行 
# zkServer.sh start 
启动后可以使用 
# zkServer.sh status 
查看状态 

5、安装hbase 
在221上修改/home/hadoop/cdh3/hbase-0.90.1-cdh3u0/conf/hbase-env.sh添加 

 export JAVA_HOME=/usr/local/jdk1.6.0_24
 export HBASE_MANAGES_ZK=false

在221上修改/home/hadoop/cdh3/hbase-0.90.1-cdh3u0/conf/hbase-site.xml添加 

 <property>
   <name>hbase.rootdir</name>
   <value>hdfs://ubuntu-1:9000/hbase</value>
 </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>ubuntu-9,ubuntu-10,ubuntu-11</value>
 </property>

在221上修改/home/hadoop/cdh3/hbase-0.90.1-cdh3u0/conf/regionservers添加 

 ubuntu-2
 ubuntu-3
 ubuntu-4
 ubuntu-5
 ubuntu-6
 ubuntu-7
 ubuntu-8

拷贝221的hbase到222-228,232 
# scp -r /home/hadoop/cdh3/hbase-0.90.1-cdh3u0/ [email protected]:/home/hadoop/cdh3/ 

启动hbase 
在221上执行 
# start-hbase.sh 
启动hbase的第二个HMaster 
在232上执行 
# hbase-daemon.sh start master 

查看Master: http://192.168.0.221:60010/master.jsp 



查看Region Server: http://192.168.0.222:60030/regionserver.jsp 



查看ZK: http://192.168.0.221:60010/zk.jsp 




6、说明 
jps查看启动进程 
221 

 JobTracker
 NameNode
 HMaster

222-228 

 HRegionServer
 DataNode
 TaskTracker

229-231 

 QuorumPeerMain

232 

 SecondaryNameNode
 HMaster

启动顺序 
1.hadoop 
2.zookeeper 
3.hbase 
4.第二个HMaster 

停止顺序 
1.第二个HMaster, kill-9删除 
2.hbase 
3.zookeeper 
4.hadoop