Hadoop+Hbase+Zookeeper安装及配置完整版(Hadoop1系列)

第一步：安装Hadoop集群

成都创新互联公司长期为成百上千家客户提供的网站建设服务，团队从业经验10年，关注不同地域、不同群体，并针对不同对象提供差异化的产品和服务；打造开放共赢平台，与合作伙伴共同营造健康的互联网生态环境。为石台企业提供专业的成都网站制作、网站建设、外贸网站建设，石台网站改版等技术服务。拥有十载丰富建站经验和众多成功案例,为您定制开发。

1、搭建环境所需介质准备

   Enterprise-R5-U4-Server-x86_64-dvd.iso

   hadoop-1.1.1.tar.gz

   jdk-6u26-linux-x64-rpm.bin

2、创建5个节点的虚拟机

   192.168.0.202  hd202  #NameNode

   192.168.0.203  hd203  #SecondaryNameNode

   192.168.0.204  hd204  #DataNode

   192.168.0.205  hd205  #DataNode

   192.168.0.206  hd206  #DataNode

   虚拟机安装过程中，需要将sshd服务安装上。如果磁盘空间允许的话，尽可能的将系统包安装齐全了。

3、在五个节点的虚拟机中都安装Jdk(以root用户安装)

   [root@hd202 ~]# mkdir /usr/java

   [root@hd202 ~]# mv jdk-6u26-linux-x64-rpm.bin /usr/java

   [root@hd202 ~]# cd /usr/java

   [root@hd202 java]# chmod 744 jdk-6u26-linux-x64-rpm.bin

   [root@hd202 java]# ./jdk-6u26-linux-x64-rpm.bin

   [root@hd202 java]# ln -s jdk1.6.0_26 default

4、创建hadoop管理用户(5台虚拟机中都要创建用户)

   [root@hd202 ~]# useradd cbcloud   #在没有先创建用户组的情况下，直接新增用户，用户默认所属的组和用户名相同。即cbcloud.cbcloud

   [root@hd202 ~]# passwd cbcloud    #修改用户cbcloud的密码，测试环境可设置为111111

5、编辑/etc/hosts文件(使用root用户分别在五台虚拟机上都编辑)

   # Do not remove the following line, or various programs

   # that require network functionality will fail.

   127.0.0.1       localhost.localdomain localhost

   ::1             localhost6.localdomain6 localhost6

   192.168.0.202   hd202

   192.168.0.203   hd203

   192.168.0.204   hd204

   192.168.0.205   hd205

   192.168.0.206   hd206

6、编辑/etc/sysconfig/network文件(使用root用户分别在五台虚拟机上都编辑)

   NETWORKING=yes

   NETWORKING_IPV6=no

   HOSTNAME=hd202      #主机名(192.168.0.203上应该改为hd203，以此类推，五台机器都要修改为相应的名称)

   GATEWAY=192.168.0.1

7、在五台机器之间配置用户等价性(以前面创建的用户cbcloud登陆进行操作)

   [cbcloud@hd202 ~]$ mkdir .ssh

   [cbcloud@hd202 ~]$ chmod 700 .ssh

   [cbcloud@hd202 ~]$ ssh-keygen -t rsa

   [cbcloud@hd202 ~]$ ssh-keygen -t dsa

   [cbcloud@hd202 ~]$ cat ~/.ssh/id_rsa.pub > ~/.ssh/authorized_keys

   [cbcloud@hd202 ~]$ cat ~/.ssh/id_dsa.pub > ~/.ssh/authorized_keys

   [cbcloud@hd203 ~]$ mkdir .ssh

   [cbcloud@hd203 ~]$ chmod 700 .ssh

   [cbcloud@hd203 ~]$ ssh-keygen -t rsa

   [cbcloud@hd203 ~]$ ssh-keygen -t dsa

   [cbcloud@hd203 ~]$ cat ~/.ssh/id_rsa.pub > ~/.ssh/authorized_keys

   [cbcloud@hd203 ~]$ cat ~/.ssh/id_dsa.pub > ~/.ssh/authorized_keys

   [cbcloud@hd204 ~]$ mkdir .ssh

   [cbcloud@hd204 ~]$ chmod 700 .ssh

   [cbcloud@hd204 ~]$ ssh-keygen -t rsa

   [cbcloud@hd204 ~]$ ssh-keygen -t dsa

   [cbcloud@hd204 ~]$ cat ~/.ssh/id_rsa.pub > ~/.ssh/authorized_keys

   [cbcloud@hd204 ~]$ cat ~/.ssh/id_dsa.pub > ~/.ssh/authorized_keys

   [cbcloud@hd205 ~]$ mkdir .ssh

   [cbcloud@hd205 ~]$ chmod 700 .ssh

   [cbcloud@hd205 ~]$ ssh-keygen -t rsa

   [cbcloud@hd205 ~]$ ssh-keygen -t dsa

   [cbcloud@hd205 ~]$ cat ~/.ssh/id_rsa.pub > ~/.ssh/authorized_keys

   [cbcloud@hd205 ~]$ cat ~/.ssh/id_dsa.pub > ~/.ssh/authorized_keys

   [cbcloud@hd206 ~]$ mkdir .ssh

   [cbcloud@hd206 ~]$ chmod 700 .ssh

   [cbcloud@hd206 ~]$ ssh-keygen -t rsa

   [cbcloud@hd206 ~]$ ssh-keygen -t dsa

   [cbcloud@hd206 ~]$ cat ~/.ssh/id_rsa.pub > ~/.ssh/authorized_keys

   [cbcloud@hd206 ~]$ cat ~/.ssh/id_dsa.pub > ~/.ssh/authorized_keys

   [cbcloud@hd202 ~]$ cd .ssh

   [cbcloud@hd202 .ssh]$ scp authorized_keys  cbcloud@hd203:/home/cbcloud/.ssh/authorized_keys2  #将hd202机器上的authorized_keys文件远程复制到hd203上的/home/cbcloud/.ssh/目录下，并重命名为authorized_keys2

   [cbcloud@hd203 ~]$ cd .ssh

   [cbcloud@hd203 ~]$ cat authorized_keys2 > authorized_keys  #也就是将hd202上的authorized_keys中的内容合并到hd203机器上的authorized_keys文件中。

    然后再将合并后的authorized_keys文件复制到hd204上，与204上的authorized_keys文件合并，依次类推，最后将5个节点的authorized_keys文件的内容都合并在一起以后，再将包含有五个节点密钥内容的authorized_keys文件，覆盖到其余4个节点上。

    注意：authorized_keys文件的权限必须为644，否则用户等价性会失效。  

    在五个节点上都执行以下命令：

   [cbcloud@hd202 ~]$ cd .ssh

   [cbcloud@hd202 ~]$ chmod 644 authorized_keys

8、开始安装hadoop集群

8.1 建立目录 (在五台虚拟机上都执行以下命令_使用root用户)

    [root@hd202 ~]# mkdir /home/cbcloud/hdtmp

    [root@hd202 ~]# mkdir /home/cbcloud/hddata

    [root@hd202 ~]# mkdir /home/cbcloud/hdconf

    [root@hd202 ~]# chown -R cbcloud:cbcloud /home/cbcloud/hdtmp

    [root@hd202 ~]# chown -R cbcloud:cbcloud /home/cbcloud/hddata

    [root@hd202 ~]# chown -R cbcloud:cbcloud /home/cbcloud/hdconf

    [root@hd202 ~]# chmod -R 755 /home/cbcloud/hddata  #切记，hddata是用于DataNode节点存放数据用的，hadoop严格归定，这个目录的权限必须为755。如果不是这个权限值，则在后面启动DataNode时，将会因为权限不对，而不能成功启动DataNode节点。

8.2 解压hadoop-1.1.1.tar.gz到/home/cbcloud目录下(只需要在hd202一台机器上执行即可)

    [root@hd202 ~]# mv hadoop-1.1.1.tar.gz /home/cbcloud

    [root@hd202 ~]# cd /home/cbcloud

    [root@hd202 cbcloud]# tar -xzvf hadoop-1.1.1.tar.gz

    [root@hd202 cbcloud]# mv hadoop-1.1.1 hadoop

    [root@hd202 cbcloud]# chown -R cbcloud.cbcloud hadoop/

8.3 配置系统环境变量/etc/profile(在五台虚拟机上都执行_使用root用户)

    [root@hd202 ~]# vi /etc/profile

    在文件尾部加入以下内容

    export JAVA_HOME=/usr/java/default

    export CLASSPATH=$CLASSPATH:$JAVA_HOME/lib:$JAVA_HOME/jre/lib

    export PATH=$JAVA_HOME/bin:$JAVA_HOME/lib:$JAVA_HOME/jre/bin:$PATH:$HOME/bin

    export HADOOP_HOME=/home/cbcloud/hadoop

    export HADOOP_DEV_HOME=/home/cbcloud/hadoop 

    export HADOOP_COMMON_HOME=/home/cbcloud/hadoop 

    export HADOOP_HDFS_HOME=/home/cbcloud/hadoop 

    export HADOOP_CONF_DIR=/home/cbcloud/hdconf

    export HADOOP_HOME_WARN_SUPPRESS=1

    export PATH=$PATH:$HADOOP_HOME/bin

    export CLASSPATH=$CLASSPATH:$HADOOP_HOME/lib  

8.4 配置用户环境变量

    [cbcloud@hd202 ~]$ vi .bash_profile

    在文件尾部加入以下内容

    export JAVA_HOME=/usr/java/default

    export CLASSPATH=$CLASSPATH:$JAVA_HOME/lib:$JAVA_HOME/jre/lib

    export PATH=$JAVA_HOME/bin:$JAVA_HOME/lib:$JAVA_HOME/jre/bin:$PATH:$HOME/bin

    export HADOOP_HOME=/home/cbcloud/hadoop

    export HADOOP_DEV_HOME=/home/cbcloud/hadoop 

    export HADOOP_COMMON_HOME=/home/cbcloud/hadoop 

    export HADOOP_HDFS_HOME=/home/cbcloud/hadoop 

    export HADOOP_CONF_DIR=/home/cbcloud/hdconf

    export HADOOP_HOME_WARN_SUPPRESS=1

    export PATH=$PATH:$HADOOP_HOME/bin

    export CLASSPATH=$CLASSPATH:$HADOOP_HOME/lib  

8.5 修改hadoop配置文件(使用cbcloud用户操作，并且只需要在hd202一台机器上操作) 

    [cbcloud@hd202 ~]$ cp $HADOOP_HOME/conf/*   $HADOOP_CONF_DIR/*  

    #从上一步的环境变量红色那一行可以看到，目前hadoop使用的配置文件应该位于/home/cbcloud/hdconf目录中，所以需要将/home/cbcloud/hadoop/conf目录下的所有配置文件都复制一份到/home/cbcloud/hdconf目录下。

8.5.1  编辑core-site.xml配置文件

    [cbcloud@hd202 ~]$ cd /home/cbcloud/hdconf

    [cbcloud@hd202 hdconf]$ vi core-site.xml

        fs.default.name

        hdfs://hd202:9000

        hadoop.tmp.dir

        /home/cbcloud/hdtmp

8.5.2 编辑hdfs-site.xml

    [cbcloud@hd202 hdconf]$ vi hdfs-site.xml

        dfs.data.dir

        /home/cbcloud/hddata

        dfs.replication

        3

8.5.3 编辑mapred-site.xml

    [cbcloud@hd202 hdconf]$ vi mapred-site.xml

        mapred.job.tracker

        hd202:9001

8.5.4 编辑masters

    [cbcloud@hd202 hdconf]$ vi masters

    加入以下内容

    hd203   # 因为hd203为SecondaryNameNode，所以在此只需要配置hd203即可，不需要配置hd202

8.5.5 编辑slaves

    [cbcloud@hd202 hdconf]$ vi slaves

    加入以下内容

    hd204

    hd205

    hd206

8.6 复制/home/cbcloud/hadoop目录和/home/cbcloud/hdconf目录到其他四台虚拟机上

    [cbcloud@hd202 hdconf]$ scp -r /home/cbcloud/hadoop hd203:/home/cbcloud  #由于前面配置了用户等价性，因此这条命令执行时不再需要密码

    [cbcloud@hd202 hdconf]$ scp -r /home/cbcloud/hadoop hd204:/home/cbcloud

    [cbcloud@hd202 hdconf]$ scp -r /home/cbcloud/hadoop hd205:/home/cbcloud

    [cbcloud@hd202 hdconf]$ scp -r /home/cbcloud/hadoop hd206:/home/cbcloud

    [cbcloud@hd202 hdconf]$ scp -r /home/cbcloud/hdconf hd203:/home/cbcloud

    [cbcloud@hd202 hdconf]$ scp -r /home/cbcloud/hdconf hd204:/home/cbcloud

    [cbcloud@hd202 hdconf]$ scp -r /home/cbcloud/hdconf hd205:/home/cbcloud

    [cbcloud@hd202 hdconf]$ scp -r /home/cbcloud/hdconf hd206:/home/cbcloud   

8.7 在NameNode(hd202)上执行命令格式化命令空间

    [cbcloud@hd202 ~]$ cd $HADOOP_HOME/bin

    [cbcloud@hd202 bin]$ hadoop namenode -format

    如果控制台打印的信息中没有ERROR之灰的信息，表示格式化命名空间命令就执行成功了。

8.8 启动hadoop

    [cbcloud@hd202 ~]$ cd $HADOOP_HOME/bin

    [cbcloud@hd202 bin]$ ./start-dfs.sh

    starting namenode, logging to /home/cbcloud/hadoop/libexec/../logs/hadoop-cbcloud-namenode-hd202.out

    hd204: starting datanode, logging to /home/cbcloud/hadoop/libexec/../logs/hadoop-cbcloud-datanode-hd204.out

    hd205: starting datanode, logging to /home/cbcloud/hadoop/libexec/../logs/hadoop-cbcloud-datanode-hd205.out

    hd206: starting datanode, logging to /home/cbcloud/hadoop/libexec/../logs/hadoop-cbcloud-datanode-hd206.out

    hd203: starting secondarynamenode, logging to /home/cbcloud/hadoop/libexec/../logs/hadoop-cbcloud-secondarynamenode-hd203.out 

8.9 启动mapred

    [cbcloud@hd202 bin]$ ./start-mapred.sh 

    starting jobtracker, logging to /home/cbcloud/hadoop/libexec/../logs/hadoop-cbcloud-jobtracker-hd202.out

    hd204: starting tasktracker, logging to /home/cbcloud/hadoop/libexec/../logs/hadoop-cbcloud-tasktracker-hd204.out

    hd205: starting tasktracker, logging to /home/cbcloud/hadoop/libexec/../logs/hadoop-cbcloud-tasktracker-hd205.out

    hd206: starting tasktracker, logging to /home/cbcloud/hadoop/libexec/../logs/hadoop-cbcloud-tasktracker-hd206.out 

8.10 查看进程

    [cbcloud@hd202 bin]$ jps

    4335 JobTracker

    4460 Jps

    4153 NameNode 

    [cbcloud@hd203 hdconf]$ jps

    1142 Jps

    1078 SecondaryNameNode

    [cbcloud@hd204 hdconf]$ jps

    1783 Jps

    1575 DataNode

    1706 TaskTracker

    [cbcloud@hd205 hdconf]$ jps

    1669 Jps

    1461 DataNode

    1590 TaskTracker

    [cbcloud@hd206 hdconf]$ jps

    1494 DataNode

    1614 TaskTracker

    1694 Jps

8.11 查看集群状态

    [cbcloud@hd202 bin]$ hadoop dfsadmin -report

    Configured Capacity: 27702829056 (25.8 GB)

    Present Capacity: 13044953088 (12.15 GB)

    DFS Remaining: 13044830208 (12.15 GB)

    DFS Used: 122880 (120 KB)

    DFS Used%: 0%

    Under replicated blocks: 0

    Blocks with corrupt replicas: 0

    Missing blocks: 0

    -------------------------------------------------

    Datanodes available: 3 (3 total, 0 dead)

    Name: 192.168.0.205:50010

    Decommission Status : Normal

    Configured Capacity: 9234276352 (8.6 GB)

    DFS Used: 40960 (40 KB)

    Non DFS Used: 4885942272 (4.55 GB)

    DFS Remaining: 4348293120(4.05 GB)

    DFS Used%: 0%

    DFS Remaining%: 47.09%

    Last contact: Wed Jan 30 18:02:17 CST 2013

    Name: 192.168.0.206:50010

    Decommission Status : Normal

    Configured Capacity: 9234276352 (8.6 GB)

    DFS Used: 40960 (40 KB)

    Non DFS Used: 4885946368 (4.55 GB)

    DFS Remaining: 4348289024(4.05 GB)

    DFS Used%: 0%

    DFS Remaining%: 47.09%

    Last contact: Wed Jan 30 18:02:17 CST 2013

    Name: 192.168.0.204:50010

    Decommission Status : Normal

    Configured Capacity: 9234276352 (8.6 GB)

    DFS Used: 40960 (40 KB)

    Non DFS Used: 4885987328 (4.55 GB)

    DFS Remaining: 4348248064(4.05 GB)

    DFS Used%: 0%

    DFS Remaining%: 47.09%

    Last contact: Wed Jan 30 18:02:17 CST 2013

注意：如果报错“INFO ipc.Client: Retrying connect to server”,是因为core-site.xml失效的原因。停止，重启hadoop后，格式化namenode即可。

另外，每次启动VM都要关闭防火墙。

8.12 通过WEB浏览器查看Hadoop运行情况

http://192.168.1.202:50070 查看Hadoop运行情况

8.13 通过WEB浏览器查看Job运行情况

http://192.168.0.202:50030 查看Job执行情况

9、列出HDFS文件系统中存在的目录情况

[cbcloud@hd202 logs]$ hadoop dfs -ls

ls: Cannot access .: No such file or directory.

上面的错误是因为被访问目录为空所致。

可以改为执行hadoop fs -ls /

[cbcloud@hd202 logs]$ hadoop fs -ls /

Found 1 items

drwxr-xr-x   - cbcloud supergroup          0 2013-01-30 15:52 /home

可以看到有一条空结果

执行hadoop fs -mkdir hello  #hello为文件夹的名字

[cbcloud@hd202 logs]$ hadoop fs -mkdir hello

[cbcloud@hd202 logs]$ hadoop fs -ls

Found 1 items

drwxr-xr-x   - cbcloud supergroup          0 2013-01-30 21:16 /user/cbcloud/hello

10、HDFS使用测试

[cbcloud@hd202 logs]$ hadoop dfs -rmr hello

Deleted hdfs://hd202:9000/user/cbcloud/hello    #删除前面创建的文件夹

[cbcloud@hd202 logs]$ hadoop dfs -mkdir input

[cbcloud@hd202 logs]$ hadoop dfs -ls

Found 1 items

drwxr-xr-x   - cbcloud supergroup          0 2013-01-30 21:18 /user/cbcloud/input

11、运行Hadoop自带框架的wordcount示例

11.1、建立数据文件

在主机192.168.0.202虚拟机中建立两个文件input1和input2

[cbcloud@hd202 hadoop]$ echo "Hello Hadoop in input1" > input1

[cbcloud@hd202 hadoop]$ echo "Hello Hadoop in input2" > input2

11.2、发布数据文件至Hadoop集群上

1、在HDFS中建立一个input目录

[cbcloud@hd202 hadoop]$ hadoop dfs -mkdir input

2、将文件input1和input2拷贝到HDFS的input目录下

[cbcloud@hd202 hadoop]$ hadoop dfs -copyFromLocal /home/cbcloud/hadoop/input* input

3、查看input目录下有没有复制成功

[cbcloud@hd202 hadoop]$ hadoop dfs -ls input

Found 2 items

-rw-r--r--   3 cbcloud supergroup         23 2013-01-30 21:28 /user/cbcloud/input/input1

-rw-r--r--   3 cbcloud supergroup         23 2013-01-30 21:28 /user/cbcloud/input/input2

11.3、执行wordcount程序  #确保HDFS上没有output目录，查看结果

[cbcloud@hd202 hadoop]$ hadoop jar hadoop-examples-1.1.1.jar wordcount input output

13/01/30 21:33:05 INFO input.FileInputFormat: Total input paths to process : 2

13/01/30 21:33:05 INFO util.NativeCodeLoader: Loaded the native-hadoop library

13/01/30 21:33:05 WARN snappy.LoadSnappy: Snappy native library not loaded

13/01/30 21:33:07 INFO mapred.JobClient: Running job: job_201301302110_0001

13/01/30 21:33:08 INFO mapred.JobClient:  map 0% reduce 0%

13/01/30 21:33:32 INFO mapred.JobClient:  map 50% reduce 0%

13/01/30 21:33:33 INFO mapred.JobClient:  map 100% reduce 0%

13/01/30 21:33:46 INFO mapred.JobClient:  map 100% reduce 100%

13/01/30 21:33:53 INFO mapred.JobClient: Job complete: job_201301302110_0001

13/01/30 21:33:53 INFO mapred.JobClient: Counters: 29

13/01/30 21:33:53 INFO mapred.JobClient:   Job Counters 

13/01/30 21:33:53 INFO mapred.JobClient:     Launched reduce tasks=1

13/01/30 21:33:53 INFO mapred.JobClient:     SLOTS_MILLIS_MAPS=29766

13/01/30 21:33:53 INFO mapred.JobClient:     Total time spent by all reduces waiting after reserving slots (ms)=0

13/01/30 21:33:53 INFO mapred.JobClient:     Total time spent by all maps waiting after reserving slots (ms)=0

13/01/30 21:33:53 INFO mapred.JobClient:     Launched map tasks=2

13/01/30 21:33:53 INFO mapred.JobClient:     Data-local map tasks=2

13/01/30 21:33:53 INFO mapred.JobClient:     SLOTS_MILLIS_REDUCES=13784

13/01/30 21:33:53 INFO mapred.JobClient:     File Output Format Counters 

13/01/30 21:33:53 INFO mapred.JobClient:     Bytes Written=40

13/01/30 21:33:53 INFO mapred.JobClient:     FileSystemCounters

13/01/30 21:33:53 INFO mapred.JobClient:     FILE_BYTES_READ=100

13/01/30 21:33:53 INFO mapred.JobClient:     HDFS_BYTES_READ=262

13/01/30 21:33:53 INFO mapred.JobClient:     FILE_BYTES_WRITTEN=71911

13/01/30 21:33:53 INFO mapred.JobClient:     HDFS_BYTES_WRITTEN=40

13/01/30 21:33:53 INFO mapred.JobClient:     File Input Format Counters 

13/01/30 21:33:53 INFO mapred.JobClient:     Bytes Read=46

13/01/30 21:33:53 INFO mapred.JobClient:     Map-Reduce Framework

13/01/30 21:33:53 INFO mapred.JobClient:     Map output materialized bytes=106

13/01/30 21:33:53 INFO mapred.JobClient:     Map input records=2

13/01/30 21:33:53 INFO mapred.JobClient:     Reduce shuffle bytes=106

13/01/30 21:33:53 INFO mapred.JobClient:     Spilled Records=16

13/01/30 21:33:53 INFO mapred.JobClient:     Map output bytes=78

13/01/30 21:33:53 INFO mapred.JobClient:     CPU time spent (ms)=5500

13/01/30 21:33:53 INFO mapred.JobClient:     Total committed heap usage (bytes)=336928768

13/01/30 21:33:53 INFO mapred.JobClient:     Combine input records=8

13/01/30 21:33:53 INFO mapred.JobClient:     SPLIT_RAW_BYTES=216

13/01/30 21:33:53 INFO mapred.JobClient:     Reduce input records=8

13/01/30 21:33:53 INFO mapred.JobClient:     Reduce input groups=5

13/01/30 21:33:53 INFO mapred.JobClient:     Combine output records=8

13/01/30 21:33:53 INFO mapred.JobClient:     Physical memory (bytes) snapshot=417046528

13/01/30 21:33:53 INFO mapred.JobClient:     Reduce output records=5

13/01/30 21:33:53 INFO mapred.JobClient:     Virtual memory (bytes) snapshot=1612316672

13/01/30 21:33:53 INFO mapred.JobClient:     Map output records=8

[cbcloud@hd202 hadoop]$ hadoop dfs -ls output

Found 2 items

-rw-r--r--   3 cbcloud supergroup          0 2013-01-30 21:33 /user/cbcloud/output/_SUCCESS

-rw-r--r--   3 cbcloud supergroup         40 2013-01-30 21:33 /user/cbcloud/output/part-r-00000

[cbcloud@hd202 hadoop]$ hadoop dfs -cat output/part-r-00000

Hadoop  2

Hello   2

in      2

input1  1

input2  1

第二步：搭建Zookeeper集群环境

上一篇关于Hadoop1.1.1集群安装记录中已经详细记录了在Oracle Linux 5.4 64bit上搭建Hadoop集群的方法。现在接着上一篇的内容，进一步安装Zookeeper和HBASE

1、安装zookeeper (在hd202上安装)

1.1、准备安装介质zookeeper-3.4.5.tar.gz

1.2、使用cbcloud用户将介质上传到hd202虚拟机上的/home/cbcloud/目录下面

1.3、解压缩zookeeper-3.4.5.tar.gz

[cbcloud@hd202 ~]$ tar zxvf zookeeper-3.4.5.tar.gz

1.4、在hd204、hd205、hd206三台机器上创建目录

[cbcloud@hd204 ~]$ mkdir /home/cbcloud/zookeeperdata

[cbcloud@hd205 ~]$ mkdir /home/cbcloud/zookeeperdata

[cbcloud@hd206 ~]$ mkdir /home/cbcloud/zookeeperdata

1.5、在hd202上执行以下内容

[cbcloud@hd202 ~]$ mv zookeeper-3.4.5 zookeeper

[cbcloud@hd202 ~]$ cd zookeeper/conf

[cbcloud@hd202 ~]$ mv zoo_sample.cfg zoo.cfg

[cbcloud@hd202 ~]$ vi 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.

# do not use /tmp for storage, /tmp here is just

# example sakes.

dataDir=/home/cbcloud/zookeeperdata

# the port at which the clients will connect

clientPort=2181

server.1=hd204:2888:3888

server.2=hd205:2888:3888

server.3=hd206:2888:3888

#

# Be sure to read the maintenance section of the

# administrator guide before turning on autopurge.

#

# http://zookeeper.apache.org/doc/current/zookeeperAdmin.html#sc_maintenance

#

# The number of snapshots to retain in dataDir

#autopurge.snapRetainCount=3

# Purge task interval in hours

# Set to "0" to disable auto purge feature

#autopurge.purgeInterval=1

1.6、将zookeeper文件夹复制到hd204、hd205、hd206三台虚拟机上

[cbcloud@hd202 ~]$ scp -r zookeeper hd204:/home/cbcloud/

[cbcloud@hd202 ~]$ scp -r zookeeper hd205:/home/cbcloud/

[cbcloud@hd202 ~]$ scp -r zookeeper hd206:/home/cbcloud/

1.7、在hd204、hd205、hd206三台虚拟机的/home/cbcloud/zookeeperdata目录下新建一个myid文件，并依次插入数字1、2、3

[cbcloud@hd204 ~]$ cd zookeeperdata

[cbcloud@hd204 zookeeperdata]$ touch myid

[cbcloud@hd204 zookeeperdata]$ vi myid

加入以下内容

1          #与前面配置文件中的server.1=hd204:2888:3888的编号相对应

[cbcloud@hd205 ~]$ cd zookeeperdata

[cbcloud@hd205 zookeeperdata]$ touch myid

[cbcloud@hd205 zookeeperdata]$ vi myid

加入以下内容

2          #与前面配置文件中的server.2=hd205:2888:3888的编号相对应

[cbcloud@hd206 ~]$ cd zookeeperdata

[cbcloud@hd206 zookeeperdata]$ touch myid

[cbcloud@hd206 zookeeperdata]$ vi myid

加入以下内容

3          #与前面配置文件中的server.3=hd206:2888:3888的编号相对应

1.8、启动zookeeper，在hd204、hd205、hd206机器上的/home/cbcloud/zookeeper/bin目录下执行zkServer.sh start

[cbcloud@hd204 ~]$ cd zookeeper

[cbcloud@hd204 zookeeper]$ cd bin

[cbcloud@hd204 bin]$ ./zkServer.sh start

JMX enabled by default

Using config: /home/cbcloud/zookeeper/bin/../conf/zoo.cfg

Starting zookeeper ... STARTED

[cbcloud@hd205 ~]$ cd zookeeper

[cbcloud@hd205 zookeeper]$ cd bin

[cbcloud@hd205 bin]$ ./zkServer.sh start

JMX enabled by default

Using config: /home/cbcloud/zookeeper/bin/../conf/zoo.cfg

Starting zookeeper ... STARTED

[cbcloud@hd206 ~]$ cd zookeeper

[cbcloud@hd206 zookeeper]$ cd bin

[cbcloud@hd206 bin]$ ./zkServer.sh start

JMX enabled by default

Using config: /home/cbcloud/zookeeper/bin/../conf/zoo.cfg

Starting zookeeper ... STARTED

1.9 查看zookeeper的进程状态

[cbcloud@hd204 bin]$ ./zkServer.sh status

JMX enabled by default

Using config: /home/cbcloud/zookeeper/bin/../conf/zoo.cfg

Mode: follower   #从此模式可以看出，hd204当前为跟随者模式

[cbcloud@hd205 bin]$ ./zkServer.sh status

JMX enabled by default

Using config: /home/cbcloud/zookeeper/bin/../conf/zoo.cfg

Mode: leader  #从此模式可以看出，hd204当前为领导模式

[cbcloud@hd206 bin]$ ./zkServer.sh status

JMX enabled by default

Using config: /home/cbcloud/zookeeper/bin/../conf/zoo.cfg

Mode: follower   #从此模式可以看出，hd206当前为跟随者模式

2、查看 zookeeper的进程详细状态

[cbcloud@hd204 bin]$ echo stat |nc localhost 2181

Zookeeper version: 3.4.5-1392090, built on 09/30/2012 17:52 GMT

Clients:

 /127.0.0.1:41205[0](queued=0,recved=1,sent=0)

Latency min/avg/max: 0/0/0

Received: 2

Sent: 1

Connections: 1

Outstanding: 0

Zxid: 0x0

Mode: follower

Node count: 4

[cbcloud@hd205 bin]$ echo stat |nc localhost 2181

Zookeeper version: 3.4.5-1392090, built on 09/30/2012 17:52 GMT

Clients:

 /127.0.0.1:38712[0](queued=0,recved=1,sent=0)

Latency min/avg/max: 0/0/0

Received: 2

Sent: 1

Connections: 1

Outstanding: 0

Zxid: 0x100000000

Mode: leader

Node count: 4

[cbcloud@hd206 bin]$ echo stat |nc localhost 2181

Zookeeper version: 3.4.5-1392090, built on 09/30/2012 17:52 GMT

Clients:

 /127.0.0.1:39268[0](queued=0,recved=1,sent=0)

Latency min/avg/max: 0/0/0

Received: 2

Sent: 1

Connections: 1

Outstanding: 0

Zxid: 0x100000000

Mode: follower

Node count: 4

第三步：搭建HBase集群

1、准备安装介质  hbase-0.94.4.tar.gz

2、使用用户cbcloud将安装介质上传到hd202虚拟机上的/home/cbcloud/目录下

3、使用cbcloud用户登陆到hd202虚拟机上，解压缩hbase-0.94.4.tar.gz

[cbcloud@hd202 ~]$ tar zxvf hbase-0.94.4.tar.gz

[cbcloud@hd202 ~]$ mv hbase-0.94.4 hbas

4、在五台虚拟机上都创建hbase的配置文件目录hbconf (使用cbcloud用户操作)

[cbcloud@hd202 ~]$ mkdir /home/cbcloud/hbconf

[cbcloud@hd203 ~]$ mkdir /home/cbcloud/hbconf

[cbcloud@hd204 ~]$ mkdir /home/cbcloud/hbconf

[cbcloud@hd205 ~]$ mkdir /home/cbcloud/hbconf

[cbcloud@hd206 ~]$ mkdir /home/cbcloud/hbconf

5、配置系统环境变量(以root用户操作)

[root@hd202 ~]# vi /etc/profile

在文件尾部加入以下内容

export HBASE_CONF_DIR=/home/cbcloud/hbconf

export HBASE_HOME=/home/cbcloud/hbase

[root@hd203 ~]# vi /etc/profile

在文件尾部加入以下内容

export HBASE_CONF_DIR=/home/cbcloud/hbconf

export HBASE_HOME=/home/cbcloud/hbase

[root@hd204 ~]# vi /etc/profile

在文件尾部加入以下内容

export HBASE_CONF_DIR=/home/cbcloud/hbconf

export HBASE_HOME=/home/cbcloud/hbase

[root@hd205 ~]# vi /etc/profile

在文件尾部加入以下内容

export HBASE_CONF_DIR=/home/cbcloud/hbconf

export HBASE_HOME=/home/cbcloud/hbase

[root@hd206 ~]# vi /etc/profile

在文件尾部加入以下内容

export HBASE_CONF_DIR=/home/cbcloud/hbconf

export HBASE_HOME=/home/cbcloud/hbase

6、配置用户环境变量(以cbcloud用户操作)

[cbcloud@hd202 ~]$ vi .bash_profile

在文件尾部加入以下内容

export HBASE_CONF_DIR=/home/cbcloud/hbconf

export HBASE_HOME=/home/cbcloud/hbase

[cbcloud@hd203 ~]$ vi .bash_profile

在文件尾部加入以下内容

export HBASE_CONF_DIR=/home/cbcloud/hbconf

export HBASE_HOME=/home/cbcloud/hbase

[cbcloud@hd204 ~]$ vi .bash_profile

在文件尾部加入以下内容

export HBASE_CONF_DIR=/home/cbcloud/hbconf

export HBASE_HOME=/home/cbcloud/hbase

[cbcloud@hd205 ~]$ vi .bash_profile

在文件尾部加入以下内容

export HBASE_CONF_DIR=/home/cbcloud/hbconf

export HBASE_HOME=/home/cbcloud/hbase

[cbcloud@hd206 ~]$ vi .bash_profile

在文件尾部加入以下内容

export HBASE_CONF_DIR=/home/cbcloud/hbconf

export HBASE_HOME=/home/cbcloud/hbase

7、复制$HBASE_HOME目录下的conf子目录下的所有文件到$HBASE_CONF_DIR目录下(只在hd202上操作)

[cbcloud@hd202 ~]$ cp /home/cbcloud/hbase/conf/*  /home/cbcloud/hbconf/

8、编辑$HBASE_CONF_DIR目录下的hbase_env.sh(只在hd202上操作)

找到export HBASE_OPTS="-XX:+UseConcMarkSweepGC" 这一行，将其注释掉，然后添加以下内容

export HBASE_OPTS="$HBASE_OPTS -XX:+HeapDumpOnOutOfMemoryError -XX:+UseConcMarkSweepGC"

export JAVA_HOME=/usr/java/default

export HBASE_HOME=/home/cbcloud/hbase

export HADOOP_HOME=/home/cbcloud/hadoop

export HBASE_MANAGES_ZK=true    //由HBASE自动管理zookeeper进程

9、编辑$HBASE_CONF_DIR目录下的hbase_site.xml(只在hd202上操作)

   加入以下内容

         hbase.rootdir

         hdfs://hd202:9000/hbase

         hbase.cluster.distributed

         true

         hbase.master

         hd202:60000

         hbase.master.port

         60000

         The port master should bind to.

         hbase.zookeeper.quorum

         hd204,hd205,hd206

         hbase.zookeeper.property.dataDir

         /home/cbcloud/zookeeperdata

10、编辑regionservers文件

删除localhost，然后加入以下内容

hd204

hd205

hd206

11、复制$HBASE_HOME目录及$HBASE_CONF_DIR目录到其他四台虚拟机上

[cbcloud@hd202 ~]$ scp -r hbase hd203:/home/cbcloud/

[cbcloud@hd202 ~]$ scp -r hbase hd204:/home/cbcloud/

[cbcloud@hd202 ~]$ scp -r hbase hd205:/home/cbcloud/

[cbcloud@hd202 ~]$ scp -r hbase hd206:/home/cbcloud/

[cbcloud@hd202 ~]$ scp -r hbconf hd203:/home/cbcloud/

[cbcloud@hd202 ~]$ scp -r hbconf hd204:/home/cbcloud/

[cbcloud@hd202 ~]$ scp -r hbconf hd205:/home/cbcloud/

[cbcloud@hd202 ~]$ scp -r hbconf hd206:/home/cbcloud/

12、启动HBASE

[cbcloud@hd202 ~]$ cd hbase

[cbcloud@hd202 hbase]$ cd bin

[cbcloud@hd202 bin]$ ./start-hbase.sh  #在主节点上启动hbase

starting master, logging to /home/cbcloud/hbase/logs/hbase-cbcloud-master-hd202.out

hd204: starting regionserver, logging to /home/cbcloud/hbase/logs/hbase-cbcloud-regionserver-hd204.out

hd205: starting regionserver, logging to /home/cbcloud/hbase/logs/hbase-cbcloud-regionserver-hd205.out

hd206: starting regionserver, logging to /home/cbcloud/hbase/logs/hbase-cbcloud-regionserver-hd206.out

[cbcloud@hd202 bin]$ jps

3779 JobTracker

4529 HMaster

4736 Jps

3633 NameNode

[cbcloud@hd203 ~]$ cd hbase

[cbcloud@hd203 hbase]$ cd bin

[cbcloud@hd203 bin]$ ./hbase-daemon.sh start master  #在SecondaryNameNode上启动HMaster

starting master, logging to /home/cbcloud/hbase/logs/hbase-cbcloud-master-hd203.out

[cbcloud@hd203 bin]$ jps

3815 Jps

3618 SecondaryNameNode

3722 HMaster

[cbcloud@hd204 hbconf]$ jps

3690 TaskTracker

3614 DataNode

4252 Jps

3845 QuorumPeerMain

4124 HRegionServer

[cbcloud@hd205 hbconf]$ jps

3826 QuorumPeerMain

3612 DataNode

3688 TaskTracker

4085 HRegionServer

4256 Jps

[cbcloud@hd206 ~]$ jps

3825 QuorumPeerMain

3693 TaskTracker

4091 HRegionServer

4279 Jps

3617 DataNode

13、使用WEB界面查看HMaster的情况http://192.168.0.202:60010

14、关闭HBbase的方法

第一步：关闭SecondaryNameNode上的HMaster服务

[cbcloud@hd203 ~]$ cd hbase

[cbcloud@hd203 hbase]$ cd bin

[cbcloud@hd203 bin]$ ./hbase-daemon.sh stop master

stopping master.

[cbcloud@hd203 bin]$ jps

4437 Jps

3618 SecondaryNameNode

第二步：关闭NameNode上的HMaster服务

[cbcloud@hd202 ~]$ cd hbase

[cbcloud@hd202 hbase]$ cd bin

[cbcloud@hd202 bin]$ ./stop-hbase.sh 

stopping hbase...................

[cbcloud@hd202 bin]$ jps

5620 Jps

3779 JobTracker

3633 NameNode

第三步:关闭zookeeper服务

[cbcloud@hd204 ~]$ cd zookeeper/bin

[cbcloud@hd204 bin]$ ./zkServer.sh stop

JMX enabled by default

Using config: /home/cbcloud/zookeeper/bin/../conf/zoo.cfg

Stopping zookeeper ... STOPPED

[cbcloud@hd204 bin]$ jps

3690 TaskTracker

3614 DataNode

4988 Jps

[cbcloud@hd205 hbconf]$ cd ..

[cbcloud@hd205 ~]$ cd zookeeper/bin

[cbcloud@hd205 bin]$ ./zkServer.sh stop

JMX enabled by default

Using config: /home/cbcloud/zookeeper/bin/../conf/zoo.cfg

Stopping zookeeper ... STOPPED

[cbcloud@hd205 bin]$ jps

3612 DataNode

3688 TaskTracker

4920 Jps

[cbcloud@hd206 ~]$ cd zookeeper

[cbcloud@hd206 zookeeper]$ cd bin

[cbcloud@hd206 bin]$ ./zkServer.sh stop

JMX enabled by default

Using config: /home/cbcloud/zookeeper/bin/../conf/zoo.cfg

Stopping zookeeper ... STOPPED

[cbcloud@hd206 bin]$ jps

4931 Jps

3693 TaskTracker

3617 DataNode

第四步：关闭hadoop

[cbcloud@hd202 bin]$ ./stop-all.sh 

stopping jobtracker

hd205: stopping tasktracker

hd204: stopping tasktracker

hd206: stopping tasktracker

stopping namenode

hd205: stopping datanode

hd206: stopping datanode

hd204: stopping datanode

hd203: stopping secondarynamenode

15、启动HBase的顺序与上面的顺序严格相反

    第一步：启动hadoop

    第二步：启动各个DataNode节点上的zookeeper

第三步：启动NameNode上的HMaster

第四步：启动SecondaryNameNode上的HMaster