一、安装centos6.4 x86_64 虚拟机7台；
1．修改Linux主机名
hadoop01
hadoop02
hadoop03
hadoop04
hadoop05
hadoop06
hadoop07
2．修改IP
192.168.56.201
192.168.56.202
192.168.56.203
192.168.56.204
192.168.56.205
192.168.56.206
192.168.56.207
3．修改主机名和IP的映射关系
192.168.56.201 hadoop01
192.168.56.202 hadoop02
192.168.56.203 hadoop03
192.168.56.204 hadoop04
192.168.56.205 hadoop05
192.168.56.206 hadoop06
192.168.56.207 hadoop07
4．关闭防火墙
service iptables stop
chkconfig iptables off
5．ssh免密码登陆
6．安装JDK，配置环境变量等
yum install -y java-1.7.0-openjdk java-1.7.0-openjdk-devel
二、规划集群
| 主机名 | IP | 安装的软件 | 运行的进程 |
| hadoop01 | 192.168.56.201 | jdk、hadoop | NameNode、DFSZKFailoverController |
| hadoop02 | 192.168.56.202 | jdk、hadoop | NameNode、DFSZKFailoverController |
| hadoop03 | 192.168.56.203 | jdk、hadoop | ResourceManager
|
| hadoop04 | 192.168.56.204 | jdk、hadoop | ResourceManager
|
| hadoop05 | 192.168.56.205 | jdk、hadoop、zookeeper | DataNode、NodeManager、JournalNode、QuorumPeerMain |
| hadoop06 | 192.168.56.206 | jdk、hadoop、zookeeper | DataNode、NodeManager、JournalNode、QuorumPeerMain |
| hadoop07 | 192.168.56.207 | jdk、hadoop、zookeeper | DataNode、NodeManager、JournalNode、QuorumPeerMain |
说明：
1.在hadoop2.0中通常由两个NameNode组成，一个处于active状态，另一个处于standby状态。Active NameNode对外提供服务，而Standby NameNode则不对外提供服务，仅同步active namenode的状态，以便能够在它失败时快速进行切换。hadoop2.0官方提供了两种HDFS HA的解决方案，一种是NFS，另一种是QJM。这里我们使用简单的QJM。在该方案中，主备NameNode之间通过一组JournalNode同步元数据信息，一条数据只要成功写入多数JournalNode即认为写入成功。通常配置奇数个JournalNode这里还配置了一个zookeeper集群，用于ZKFC（DFSZKFailoverController）故障转移，当Active NameNode挂掉了，会自动切换Standby NameNode为standby状态。
2.hadoop-2.2.0中依然存在一个问题，就是ResourceManager只有一个，存在单点故障，hadoop-2.4.1解决了这个问题，有两个ResourceManager，一个是Active，一个是Standby，状态由zookeeper进行协调。
三、安装配置zookeeper（hadoop05上操作）
1．解压zookeeper
[hadoop@hadoop05 ~]$ tar -zxvf zookeeper-3.4.6.tar.gz -C /hadoop/
2．配置zookeeper
/hadoop/zookeeper-3.4.6/conf
[hadoop@hadoop05 conf]$ cp zoo_sample.cfg zoo.cfg
创建文件夹/hadoop/zookeeper-3.4.6/tmp
修改zoo.cfg中的dataDir参数为/hadoop/zookeeper-3.4.6/tmp
添加如下内容：
server.1=hadoop05:2888:3888
server.2=hadoop06:2888:3888
server.3=hadoop07:2888:3888
在dataDir目录下创建myid
[hadoop@hadoop05 tmp]$ touch myid
[hadoop@hadoop05 tmp]$ echo 1 > myid
3．将配置好的zookeeper拷贝到其他节点
scp -r /hadoop/zookeeper-3.4.6/ hadoop06:/hadoop/
scp -r /hadoop/zookeeper-3.4.6/ hadoop07:/hadoop/
修改06、07节点的myid
[hadoop@hadoop06 ~]$ echo 2 > /hadoop/zookeeper-3.4.6/tmp/myid
[hadoop@hadoop07 ~]$ echo 3 > /hadoop/zookeeper-3.4.6/tmp/myid
四、安装配置hadoop（hadoop01上操作）
1．解压hadoop
[hadoop@hadoop01 ~]$ tar -zxvf hadoop-2.6.0-x64.tar.gz -C /hadoop/
2．配置hdfs
注意：hadoop2.*所有的配置文件都在$HADOOP_HOME/etc/hadoop目录下
（1）配置hadoop用户环境变量
vi ~/.bash_profile
export JAVA_HOME=/usr/lib/jvm/java-1.7.0-openjdk.x86_64
export HADOOP_HOME=/hadoop/hadoop-2.6.0
export PATH=$PATH:$JAVA_HOME/bin:HADOOP_HOME/bin
（2）修改hadoop-env.sh
修改export JAVA_HOME=${JAVA_HOME}为
export JAVA_HOME=/usr/lib/jvm/java-1.7.0-openjdk.x86_64
（3）修改core-site.xml
<!-- 指定hdfs的nameservice为ns1 -->
<property>
<name>fs.defaultFS</name>
<value>hdfs://ns1/</value>
</property>
<!-- 指定hadoop临时目录，不存在请创建-->
<property>
<name>hadoop.tmp.dir</name>
<value>/hadoop/hadoop-2.6.0/tmp</value>
</property>
<!-- 指定zookeeper地址-->
<property>
<name>ha.zookeeper.quorum</name>
<value>hadoop05:2181,hadoop06:2181,hadoop07:2181</value>
</property>
</configuration>
（4）修改hdfs-site.xml
<configuration>
<!--指定hdfs的nameservice为ns1，需要和core-site.xml中的保持一致-->
<property>
<name>dfs.nameservices</name>
<value>ns1</value>
</property>
<!-- ns1下面有两个NameNode，分别是nn1，nn2 -->
<property>
<name>dfs.ha.namenodes.ns1</name>
<value>nn1,nn2</value>
</property>
<!-- nn1的RPC通信地址-->
<property>
<name>dfs.namenode.rpc-address.ns1.nn1</name>
<value>hadoop01:9000</value>
</property>
<!-- nn1的http通信地址-->
<property>
<name>dfs.namenode.http-address.ns1.nn1</name>
<value>hadoop01:50070</value>
</property>
<!-- nn2的RPC通信地址-->
<property>
<name>dfs.namenode.rpc-address.ns1.nn2</name>
<value>hadoop02:9000</value>
</property>
<!-- nn2的http通信地址-->
<property>
<name>dfs.namenode.http-address.ns1.nn2</name>
<value>hadoop02:50070</value>
</property>
<!-- 指定NameNode的元数据在JournalNode上的存放位置-->
<property>
<name>dfs.namenode.shared.edits.dir</name>
<value>qjournal://hadoop05:8485;hadoop06:8485;hadoop07:8485/ns1</value>
</property>
<!-- 指定JournalNode在本地磁盘存放数据的位置-->
<property>
<name>dfs.journalnode.edits.dir</name>
<value>/hadoop/hadoop-2.6.0/journaldata</value>
</property>
<!-- 开启NameNode失败自动切换-->
<property>
<name>dfs.ha.automatic-failover.enabled</name>
<value>true</value>
</property>
<!-- 配置失败自动切换实现方式 -->
<property>
<name>dfs.client.failover.proxy.provider.ns1</name>
<value>org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider</value>
</property>
<!-- 配置隔离机制方法，多个机制用换行分割，即每个机制暂用一行-->
<property>
<name>dfs.ha.fencing.methods</name>
<value>
sshfence
shell(/bin/true)
</value>
</property>
<!-- 使用sshfence隔离机制时需要ssh免登陆-->
<property>
<name>dfs.ha.fencing.ssh.private-key-files</name>
<value>/home/hadoop/.ssh/id_rsa</value>
</property>
<!-- 配置sshfence隔离机制超时时间-->
<property>
<name>dfs.ha.fencing.ssh.connect-timeout</name>
<value>30000</value>
</property>
</configuration>
（5）修改mapred-site.xml
<configuration>
<!-- 指定mr框架为yarn方式-->
<property>
<name>mapreduce.framework.name</name>
<value>yarn</value>
</property>
</configuration>
（6）修改yarn-site.xml
<configuration>
<!-- 开启RM高可用-->
<property>
<name>yarn.resourcemanager.ha.enabled</name>
<value>true</value>
</property>
<!-- 指定RM的cluster id -->
<property>
<name>yarn.resourcemanager.cluster-id</name>
<value>yrc</value>
</property>
<!-- 指定RM的名字-->
<property>
<name>yarn.resourcemanager.ha.rm-ids</name>
<value>rm1,rm2</value>
</property>
<!-- 分别指定RM的地址-->
<property>
<name>yarn.resourcemanager.hostname.rm1</name>
<value>hadoop03</value>
</property>
<property>
<name>yarn.resourcemanager.hostname.rm2</name>
<value>hadoop04</value>
</property>
<!-- 指定zk集群地址-->
<property>
<name>yarn.resourcemanager.zk-address</name>
<value>hadoop05:2181,hadoop06:2181,hadoop07:2181</value>
</property>
<property>
<name>yarn.nodemanager.aux-services</name>
<value>mapreduce_shuffle</value>
</property>
</configuration>
（7）修改slaves
hadoop05
hadoop06
hadoop07
3．将配置好的hadoop拷贝到其他节点
scp -r /hadoop/hadoop-2.6.0/ hadoop02:/hadoop/
scp -r /hadoop/hadoop-2.6.0/ hadoop03:/hadoop/
scp -r /hadoop/hadoop-2.6.0/ hadoop04:/hadoop/
scp -r /hadoop/hadoop-2.6.0/ hadoop05:/hadoop/
scp -r /hadoop/hadoop-2.6.0/ hadoop06:/hadoop/
scp -r /hadoop/hadoop-2.6.0/ hadoop07:/hadoop/
4．启动zookeeper集群（hadoop 05、06、07上执行）
cd /hadoop/zookeeper-3.4.6/bin
./zkServer.sh start
#查看状态：一个leader，两个follower
./zkServer.sh status
[hadoop@hadoop05 bin]$ ./zkServer.sh status
JMX enabled by default
Using config: /hadoop/zookeeper-3.4.6/bin/../conf/zoo.cfg
Mode: follower
[hadoop@hadoop06 bin]$ ./zkServer.sh status
JMX enabled by default
Using config: /hadoop/zookeeper-3.4.6/bin/../conf/zoo.cfg
Mode: leader
[hadoop@hadoop07 bin]$ ./zkServer.sh status
JMX enabled by default
Using config: /hadoop/zookeeper-3.4.6/bin/../conf/zoo.cfg
Mode: follower
5．启动journalnode（hadoop 05、06、07上执行）
cd /hadoop/hadoop-2.6.0/sbin
./hadoop-daemon.sh start journalnode
#运行jps命令检验，hadoop 05、06、07上多了JournalNode进程
[hadoop@hadoop07 sbin]$ jps
1887 JournalNode
1933 Jps
1788 QuorumPeerMain
6．格式化HDFS
hadoop01节点上执行：
hdfs namenode -format
执行完成后在将/hadoop/hadoop-2.6.0/tmp文件夹拷贝到02、03节点
[hadoop@hadoop01 tmp]$ scp -r /hadoop/hadoop-2.6.0/tmp hadoop02:/hadoop/hadoop-2.6.0/
[hadoop@hadoop01 tmp]$ scp -r /hadoop/hadoop-2.6.0/tmp hadoop03:/hadoop/hadoop-2.6.0/
7．格式化ZKFC(在hadoop01上执行即可)
hdfs zkfc -formatZK
8．启动HDFS(在hadoop01上执行)
[hadoop@hadoop01 hadoop-2.6.0]$ start-dfs.sh
Starting namenodes on [hadoop01 hadoop02]
hadoop01: starting namenode, logging to /hadoop/hadoop-2.6.0/logs/hadoop-hadoop-namenode-hadoop01.out
hadoop02: starting namenode, logging to /hadoop/hadoop-2.6.0/logs/hadoop-hadoop-namenode-hadoop02.out
hadoop05: starting datanode, logging to /hadoop/hadoop-2.6.0/logs/hadoop-hadoop-datanode-hadoop05.out
hadoop07: starting datanode, logging to /hadoop/hadoop-2.6.0/logs/hadoop-hadoop-datanode-hadoop07.out
hadoop06: starting datanode, logging to /hadoop/hadoop-2.6.0/logs/hadoop-hadoop-datanode-hadoop06.out
Starting journal nodes [hadoop05 hadoop06 hadoop07]
hadoop07: starting journalnode, logging to /hadoop/hadoop-2.6.0/logs/hadoop-hadoop-journalnode-hadoop07.out
hadoop05: starting journalnode, logging to /hadoop/hadoop-2.6.0/logs/hadoop-hadoop-journalnode-hadoop05.out
hadoop06: starting journalnode, logging to /hadoop/hadoop-2.6.0/logs/hadoop-hadoop-journalnode-hadoop06.out
Starting ZK Failover Controllers on NN hosts [hadoop01 hadoop02]
hadoop01: starting zkfc, logging to /hadoop/hadoop-2.6.0/logs/hadoop-hadoop-zkfc-hadoop01.out
hadoop02: starting zkfc, logging to /hadoop/hadoop-2.6.0/logs/hadoop-hadoop-zkfc-hadoop02.out
9．启动YARN
[hadoop@hadoop03 sbin]$ ./start-yarn.sh
starting yarn daemons
starting resourcemanager, logging to /hadoop/hadoop-2.6.0/logs/yarn-hadoop-resourcemanager-hadoop03.out
hadoop06: starting nodemanager, logging to /hadoop/hadoop-2.6.0/logs/yarn-hadoop-nodemanager-hadoop06.out
hadoop07: starting nodemanager, logging to /hadoop/hadoop-2.6.0/logs/yarn-hadoop-nodemanager-hadoop07.out
hadoop05: starting nodemanager, logging to /hadoop/hadoop-2.6.0/logs/yarn-hadoop-nodemanager-hadoop05.out注意：是在hadoop03上执行start-yarn.sh，把namenode和resourcemanager分开是因为性能问题，因为他们都要占用大量资源，所以把他们分开了，他们分开了就要分别在不同的机器上启动。
五、验证安装
1．通过浏览器访问
2．验证hdfs高可用
上传一个文件到hdfs：
[hadoop@hadoop01 ~]$ ll hadoop-2.6.0-x64.tar.gz
-rwxr-xr-x. 1 hadoop root 180197035 Jul 5 12:47 hadoop-2.6.0-x64.tar.gz
[hadoop@hadoop01 ~]$ hadoop fs -put ~/hadoop-2.6.0-x64.tar.gz /
然后在命令未完成的情况下在hadoop02上杀掉active的namenode进程
[hadoop@hadoop02 .ssh]$ jps
4376 NameNode
4995 Jps
4474 DFSZKFailoverController
[hadoop@hadoop02 .ssh]$ kill -9 4376
然后查看web页面发现hadoop01立马变为active状态
而且最终文件上传成功，未损失任何数据
[hadoop@hadoop01 ~]$ ll hadoop-2.6.0-x64.tar.gz
-rwxr-xr-x. 1 hadoop root 180197035 Jul 5 12:47 hadoop-2.6.0-x64.tar.gz
[hadoop@hadoop01 ~]$ hadoop fs -ls /
Found 1 items
-rw-r--r-- 3 hadoop supergroup 180197035 2016-08-02 10:52 /hadoop-2.6.0-x64.tar.gz
验证完成后在hadoop02上将namenode再次启动起来。
[hadoop@hadoop02 ~]$ hadoop-daemon.sh start namenode
starting namenode, logging to /hadoop/hadoop-2.6.0/logs/hadoop-hadoop-namenode-hadoop02.out
启动完成查看hadoop02，已变为standby状态
3．验证yarn
（1）创建测试数据
[hadoop@hadoop01 ~]$ cat test.txt
aaa qwer
bbb asdf
ccc zxcv
zxcv zxcv
qwer qwer
bbb aaa
aaa ccc
（2）上传测试数据到hdfs
[hadoop@hadoop01 ~]$ hadoop fs -put /home/hadoop/test.txt /
[hadoop@hadoop01 ~]$ hadoop fs -ls /
Found 3 items
-rw-r--r-- 3 hadoop supergroup 180197035 2016-08-02 10:52 /hadoop-2.6.0-x64.tar.gz
-rw-r--r-- 3 hadoop supergroup 63 2016-08-02 11:03 /test.txt
drwx------ - hadoop supergroup 0 2016-08-02 11:01 /tmp
（3）运行hadoop demo中的WordCount程序
[hadoop@hadoop01 ~]$ hadoop jar hadoop-mapreduce-examples-2.6.0.jar wordcount /test.txt /out
16/08/02 11:04:45 INFO input.FileInputFormat: Total input paths to process : 1
16/08/02 11:04:46 INFO mapreduce.JobSubmitter: number of splits:1
16/08/02 11:04:47 INFO mapreduce.JobSubmitter: Submitting tokens for job: job_1470105413367_0004
16/08/02 11:04:48 INFO impl.YarnClientImpl: Submitted application application_1470105413367_0004
16/08/02 11:04:48 INFO mapreduce.Job: The url to track the job: http://hadoop03:8088/proxy/application_1470105413367_0004/
16/08/02 11:04:48 INFO mapreduce.Job: Running job: job_1470105413367_0004
16/08/02 11:04:58 INFO mapreduce.Job: Job job_1470105413367_0004 running in uber mode : false
16/08/02 11:04:58 INFO mapreduce.Job: map 0% reduce 0%
16/08/02 11:05:07 INFO mapreduce.Job: map 100% reduce 0%
16/08/02 11:05:16 INFO mapreduce.Job: map 100% reduce 100%
16/08/02 11:05:20 INFO mapreduce.Job: Job job_1470105413367_0004 completed successfully
16/08/02 11:05:20 INFO mapreduce.Job: Counters: 49
File System Counters
FILE: Number of bytes read=69
FILE: Number of bytes written=216073
FILE: Number of read operations=0
FILE: Number of large read operations=0
FILE: Number of write operations=0
HDFS: Number of bytes read=147
HDFS: Number of bytes written=39
HDFS: Number of read operations=6
HDFS: Number of large read operations=0
HDFS: Number of write operations=2
Job Counters
Launched map tasks=1
Launched reduce tasks=1
Data-local map tasks=1
Total time spent by all maps in occupied slots (ms)=6289
Total time spent by all reduces in occupied slots (ms)=6677
Total time spent by all map tasks (ms)=6289
Total time spent by all reduce tasks (ms)=6677
Total vcore-seconds taken by all map tasks=6289
Total vcore-seconds taken by all reduce tasks=6677
Total megabyte-seconds taken by all map tasks=6439936
Total megabyte-seconds taken by all reduce tasks=6837248
Map-Reduce Framework
Map input records=7
Map output records=14
Map output bytes=119
Map output materialized bytes=69
Input split bytes=84
Combine input records=14
Combine output records=6
Reduce input groups=6
Reduce shuffle bytes=69
Reduce input records=6
Reduce output records=6
Spilled Records=12
Shuffled Maps =1
Failed Shuffles=0
Merged Map outputs=1
GC time elapsed (ms)=117
CPU time spent (ms)=1000
Physical memory (bytes) snapshot=311525376
Virtual memory (bytes) snapshot=1873289216
Total committed heap usage (bytes)=136646656
Shuffle Errors
BAD_ID=0
CONNECTION=0
IO_ERROR=0
WRONG_LENGTH=0
WRONG_MAP=0
WRONG_REDUCE=0
File Input Format Counters
Bytes Read=63
File Output Format Counters
Bytes Written=39
（4）查看输出结果
[hadoop@hadoop01 ~]$ hadoop fs -cat /out/part-r-00000
aaa 3
asdf 1
bbb 2
ccc 2
qwer 3
zxcv 3
扫一扫
3298
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
