共计 13817 个字符，预计需要花费 35 分钟才能阅读完成。
部署 Hadoop 高性能集群：
拓扑图：
一、实验前期环境准备：
1、三台主机配置 hosts 文件：（复制到另外两台主机上）
[root@tiandong63 ~]# more /etc/hosts
192.168.199.3 tiandong63
192.168.199.4 tiandong64
192.168.199.5 tiandong65
2、创建 Hadoop 账号（另外两台主机上都的创建）
[root@tiandong63 ~]#useradd -u 8000 hadoop
[root@tiandong63 ~]#echo ‘123456’ | passwd –stdin hadoop
3、给 hadoop 用户增加 sudo 权限，增加内容（另外两台主机上都得配置）
[root@tiandong63 ~]# vim /etc/sudoers
hadoop ALL=(ALL) ALL
4、主机互信（在 tiandong63 主机上）
可以 ssh 无密码登录机器 tiandong63，tiandong64，tiandong65，方便后期复制文件和启动服务。因为 namenode 启动时，会连接到 datanode 上启动对应的服务。
[root@tiandong63 ~]# su – hadoop
[hadoop@tiandong63 ~]$ ssh-keygen
[hadoop@tiandong63 ~]$ ssh-copy-id root@192.168.199.4
[hadoop@tiandong63 ~]$ ssh-copy-id root@192.168.199.5
二、配置 hadoop 环境
1、配置 jdk 环境：（三台都的配置）
[root@tiandong63 ~]# ll jdk-8u191-linux-x64.tar.gz
-rw-r–r– 1 root root 191753373 Dec 30 00:58 jdk-8u191-linux-x64.tar.gz
[root@tiandong63 ~]# tar zxvf jdk-8u191-linux-x64.tar.gz -C /usr/local/src/
[root@tiandong63 ~]# vim /etc/profile
export Java_HOME=/usr/local/src/jdk1.8.0_191
export JAVA_BIN=/usr/local/src/jdk1.8.0_191/bin
export PATH=${JAVA_HOME}/bin:$PATH
export CLASSPATH=.:${JAVA_HOME}/lib/dt.jar:${JAVA_HOME}/lib/tools.jar
export hadoop_root_logger=DEBUG,console
[root@tiandong63 ~]# source /etc/profile
2、关闭防火墙（三台都的关闭）
[root@tiandong63 ~]# /etc/init.d/iptables stop
3、在 tiandong63 安装 Hadoop 并配置成 namenode 主节点
[root@tiandong63 ~]# cd /home/hadoop/
[root@tiandong63 hadoop]# ll hadoop-2.7.7.tar.gz
-rw-r–r– 1 hadoop hadoop 218720521 Dec 29 21:11 hadoop-2.7.7.tar.gz
[root@tiandong63 ~]# su – hadoop
[hadoop@tiandong63 ~]$ tar -zxvf hadoop-2.7.7.tar.gz
创建 hadoop 相关的工作目录：
[hadoop@tiandong63 ~]$ mkdir -p /home/hadoop/dfs/name/ /home/hadoop/dfs/data/^Chome/hadoop/tmp/
4、配置 hadoop（修改 7 个配置文件）
文件名称：hadoop-env.sh、yarn-evn.sh、slaves、core-site.xml、hdfs-site.xml、mapred-site.xml、yarn-site.xml
[hadoop@tiandong63 ~]$ cd /home/hadoop/hadoop-2.7.7/etc/hadoop/
[hadoop@tiandong63 hadoop]$ vim hadoop-env.sh 指定 hadoop 的 Java 运行环境
25 export JAVA_HOME=/usr/local/src/jdk1.8.0_191
[hadoop@tiandong63 hadoop]$ vim yarn-env.sh 指定 yarn 框架的 java 运行环境
26 JAVA_HOME=/usr/local/src/jdk1.8.0_191
[hadoop@tiandong63 hadoop]$ vim slaves 指定datanode 数据存储服务器
tiandong64
tiandong65
[hadoop@tiandong63 hadoop]$ vim core-site.xml 指定访问hadoop web 界面访问路径
19 <configuration>
20 <property>
21 <name>fs.defaultFS</name>
22 <value>hdfs://tiandong63:9000</value>
23 </property>
24
25 <property>
26 <name>io.file.buffer.size</name>
27 <value>131072</value>
28 </property>
29
30 <property>
31 <name>hadoop.tmp.dir</name>
32 <value>file:/home/hadoop/tmp</value>
33 <description>Abase for other temporary directories.</description>
34 </property>
35 </configuration>
[hadoop@tiandong63 hadoop]$ mkdir -p /home/hadoop/tmp
[hadoop@tiandong63 hadoop]$ vim hdfs-site.xml
hdfs 的配置文件，dfs.http.address 配置了 hdfs 的 http 的访问位置，dfs.replication 配置了文件块的副本数，一般不大于从机的个数。
19 <configuration>
20 <property>
21 <name>dfs.namenode.secondary.http-address</name>
22 <value>tiandong63:9001</value>
23 </property>
24
25 <property>
26 <name>dfs.namenode.name.dir</name>
27 <value>file:/home/hadoop/dfs/name</value>
28 </property>
29 <property>
30 <name>dfs.datanode.data.dir</name>
31 <value>file:/home/hadoop/dfs/data</value>
32 </property>
33
34 <property>
35 <name>dfs.replication</name>
36 <value>2</value>
37 </property>
38
39 <property>
40 <name>dfs.webhdfs.enabled</name>
41 <value>true</value>
42 </property>
43 </configuration>
[hadoop@tiandong63 hadoop]$ vim mapred-site.xml
19 <configuration>
20 <property>
21 <name>mapreduce.framework.name</name>
22 <value>yarn</value>
23 </property>
24
25 <property>
26 <name>mapreduce.jobhistory.address</name>
27 <value>tiandong63:10020</value>
28 </property>
29
30 <property>
31 <name>mapreduce.jobhistory.webapp.address</name>
32 <value>tiandong:19888</value>
33 </property>
34 </configuration>
[hadoop@tiandong63 hadoop]$ vim yarn-site.xml 该文件为yarn 框架的配置, 主要是一些任务的启动位置
15 <configuration>
16
17 <!– Site specific YARN configuration properties –>
18 <property>
19 <name>yarn.nodemanager.aux-services</name>
20 <value>mapreduce_shuffle</value>
21 </property>
22
23 <property>
24 <name>yarn.nodemanager.aux-services.mapreduce.shuffle.class</name>
25 <value>org.apache.hadoop.mapred.ShuffleHandler</value>
26 </property>
27
28 <property>
29 <name>yarn.resourcemanager.address</name>
30 <value>tiandong63:8032</value>
31 </property>
32
33 <property>
34 <name>yarn.resourcemanager.scheduler.address</name>
35 <value>tiandong63:8030</value>
36 </property>
37
38 <property>
39 <name>yarn.resourcemanager.resource-tracker.address</name>
40 <value>tiandong63:8031</value>
41 </property>
42
43 <property>
44 <name>yarn.resourcemanager.admin.address</name>
45 <value>tiandong63:8033</value>
46 </property>
47
48 <property>
49 <name>yarn.resourcemanager.webapp.address</name>
50 <value>tiandong63:8088</value>
51 </property>
52 </configuration>
复制到其他 datanode 节点（192.168.199.4/5）
[hadoop@tiandong63 ~]$ scp -r /home/hadoop/hadoop-2.7.7 hadoop@tiandong64
[hadoop@tiandong63 ~]$ scp -r /home/hadoop/hadoop-2.7.7 hadoop@tiandong65
三、启动 hadoop:
在 tiandong63 上面启动（使用 hadoop 用户）
1、格式化：
[hadoop@tiandong63 ~]$ cd /home/hadoop/hadoop-2.7.7/bin/
[hadoop@tiandong63 bin]$ ./hdfs nodename -format
2、启动 hdfs：
[hadoop@tiandong63 ~]$ /home/hadoop/hadoop-2.7.7/sbin/start-dfs.sh
查看进程（tiandong63 上面查看 namenode）：
[hadoop@tiandong63 ~]$ ps -axu | grep namenode –color
Warning: bad syntax, perhaps a bogus ‘-‘? See /usr/share/doc/procps-3.2.8/FAQ
hadoop 42851 0.4 17.6 2762196 177412 ? Sl 18:13 1:26 /usr/local/src/jdk1.8.0_191/bin/java -Dproc_namenode -Xmx1000m -Djava.net.preferIPv4Stack=true -Dhadoop.log.dir=/home/hadoop/hadoop-2.7.7/logs -Dhadoop.log.file=hadoop.log -Dhadoop.home.dir=/home/hadoop/hadoop-2.7.7 -Dhadoop.id.str=hadoop -Dhadoop.root.logger=INFO,console -Djava.library.path=/home/hadoop/hadoop-2.7.7/lib/native -Dhadoop.policy.file=hadoop-policy.xml -Djava.net.preferIPv4Stack=true -Djava.net.preferIPv4Stack=true -Djava.net.preferIPv4Stack=true -Dhadoop.log.dir=/home/hadoop/hadoop-2.7.7/logs -Dhadoop.log.file=hadoop-hadoop-namenode-tiandong63.log -Dhadoop.home.dir=/home/hadoop/hadoop-2.7.7 -Dhadoop.id.str=hadoop -Dhadoop.root.logger=INFO,RFA -Djava.library.path=/home/hadoop/hadoop-2.7.7/lib/native -Dhadoop.policy.file=hadoop-policy.xml -Djava.net.preferIPv4Stack=true -Dhadoop.security.logger=INFO,RFAS -Dhdfs.audit.logger=INFO,NullAppender -Dhadoop.security.logger=INFO,RFAS -Dhdfs.audit.logger=INFO,NullAppender -Dhadoop.security.logger=INFO,RFAS -Dhdfs.audit.logger=INFO,NullAppender -Dhadoop.security.logger=INFO,RFAS org.apache.hadoop.hdfs.server.namenode.NameNode
hadoop 43046 0.2 13.3 2733336 134344 ? Sl 18:13 0:35 /usr/local/src/jdk1.8.0_191/bin/java -Dproc_secondarynamenode -Xmx1000m -Djava.net.preferIPv4Stack=true -Dhadoop.log.dir=/home/hadoop/hadoop-2.7.7/logs -Dhadoop.log.file=hadoop.log -Dhadoop.home.dir=/home/hadoop/hadoop-2.7.7 -Dhadoop.id.str=hadoop -Dhadoop.root.logger=INFO,console -Djava.library.path=/home/hadoop/hadoop-2.7.7/lib/native -Dhadoop.policy.file=hadoop-policy.xml -Djava.net.preferIPv4Stack=true -Djava.net.preferIPv4Stack=true -Djava.net.preferIPv4Stack=true -Dhadoop.log.dir=/home/hadoop/hadoop-2.7.7/logs -Dhadoop.log.file=hadoop-hadoop-secondarynamenode-tiandong63.log -Dhadoop.home.dir=/home/hadoop/hadoop-2.7.7 -Dhadoop.id.str=hadoop -Dhadoop.root.logger=INFO,RFA -Djava.library.path=/home/hadoop/hadoop-2.7.7/lib/native -Dhadoop.policy.file=hadoop-policy.xml -Djava.net.preferIPv4Stack=true -Dhadoop.security.logger=INFO,RFAS -Dhdfs.audit.logger=INFO,NullAppender -Dhadoop.security.logger=INFO,RFAS -Dhdfs.audit.logger=INFO,NullAppender -Dhadoop.security.logger=INFO,RFAS -Dhdfs.audit.logger=INFO,NullAppender -Dhadoop.security.logger=INFO,RFAS org.apache.hadoop.hdfs.server.namenode.SecondaryNameNode
在 tiandong64 和 tiandong65 上面查看进程（datanode）
[root@tiandong64 ~]# ps -aux|grep datanode
Warning: bad syntax, perhaps a bogus ‘-‘? See /usr/share/doc/procps-3.2.8/FAQ
hadoop 3938 0.3 12.2 2757576 122592 ? Sl 18:13 1:04 /usr/local/src/jdk1.8.0_191/bin/java -Dproc_datanode -Xmx1000m -Djava.net.preferIPv4Stack=true -Dhadoop.log.dir=/home/hadoop/hadoop-2.7.7/logs -Dhadoop.log.file=hadoop.log -Dhadoop.home.dir=/home/hadoop/hadoop-2.7.7 -Dhadoop.id.str=hadoop -Dhadoop.root.logger=INFO,console -Djava.library.path=/home/hadoop/hadoop-2.7.7/lib/native -Dhadoop.policy.file=hadoop-policy.xml -Djava.net.preferIPv4Stack=true -Djava.net.preferIPv4Stack=true -Djava.net.preferIPv4Stack=true -Dhadoop.log.dir=/home/hadoop/hadoop-2.7.7/logs -Dhadoop.log.file=hadoop-hadoop-datanode-tiandong64.log -Dhadoop.home.dir=/home/hadoop/hadoop-2.7.7 -Dhadoop.id.str=hadoop -Dhadoop.root.logger=INFO,RFA -Djava.library.path=/home/hadoop/hadoop-2.7.7/lib/native -Dhadoop.policy.file=hadoop-policy.xml -Djava.net.preferIPv4Stack=true -server -Dhadoop.security.logger=ERROR,RFAS -Dhadoop.security.logger=ERROR,RFAS -Dhadoop.security.logger=ERROR,RFAS -Dhadoop.security.logger=INFO,RFAS org.apache.hadoop.hdfs.server.datanode.DataNode
3、启动 yarn（在 tiandong63 上面，即启动分布式计算）：
[hadoop@tiandong63 ~]$ /home/hadoop/hadoop-2.7.7/sbin/start-yarn.sh
在 tiandong63 上面查看进程（resourcemanager 进程）
[hadoop@tiandong63 ~]$ ps -ef|grep resourcemanager –color
hadoop 43196 1 0 18:14 pts/1 00:02:55 /usr/local/src/jdk1.8.0_191/bin/java -Dproc_resourcemanager -Xmx1000m -Dhadoop.log.dir=/home/hadoop/hadoop-2.7.7/logs -Dyarn.log.dir=/home/hadoop/hadoop-2.7.7/logs -Dhadoop.log.file=yarn-hadoop-resourcemanager-tiandong63.log -Dyarn.log.file=yarn-hadoop-resourcemanager-tiandong63.log -Dyarn.home.dir= -Dyarn.id.str=hadoop -Dhadoop.root.logger=INFO,RFA -Dyarn.root.logger=INFO,RFA -Djava.library.path=/home/hadoop/hadoop-2.7.7/lib/native -Dyarn.policy.file=hadoop-policy.xml -Dhadoop.log.dir=/home/hadoop/hadoop-2.7.7/logs -Dyarn.log.dir=/home/hadoop/hadoop-2.7.7/logs -Dhadoop.log.file=yarn-hadoop-resourcemanager-tiandong63.log -Dyarn.log.file=yarn-hadoop-resourcemanager-tiandong63.log -Dyarn.home.dir=/home/hadoop/hadoop-2.7.7 -Dhadoop.home.dir=/home/hadoop/hadoop-2.7.7 -Dhadoop.root.logger=INFO,RFA -Dyarn.root.logger=INFO,RFA -Djava.library.path=/home/hadoop/hadoop-2.7.7/lib/native -classpath /home/hadoop/hadoop-2.7.7/etc/hadoop:/home/hadoop/hadoop-2.7.7/etc/hadoop:/home/hadoop/hadoop-2.7.7/etc/hadoop:/home/hadoop/hadoop-2.7.7/share/hadoop/common/lib/*:/home/hadoop/hadoop-2.7.7/share/hadoop/common/*:/home/hadoop/hadoop-2.7.7/share/hadoop/hdfs:/home/hadoop/hadoop-2.7.7/share/hadoop/hdfs/lib/*:/home/hadoop/hadoop-2.7.7/share/hadoop/hdfs/*:/home/hadoop/hadoop-2.7.7/share/hadoop/yarn/lib/*:/home/hadoop/hadoop-2.7.7/share/hadoop/yarn/*:/home/hadoop/hadoop-2.7.7/share/hadoop/mapreduce/lib/*:/home/hadoop/hadoop-2.7.7/share/hadoop/mapreduce/*:/contrib/capacity-scheduler/*.jar:/contrib/capacity-scheduler/*.jar:/contrib/capacity-scheduler/*.jar:/contrib/capacity-scheduler/*.jar:/home/hadoop/hadoop-2.7.7/share/hadoop/yarn/*:/home/hadoop/hadoop-2.7.7/share/hadoop/yarn/lib/*:/home/hadoop/hadoop-2.7.7/etc/hadoop/rm-config/log4j.properties org.apache.hadoop.yarn.server.resourcemanager.ResourceManager
在 tiandong64 和 tiandong65 上面查看进程（nodemanager）
[root@tiandong64 ~]# ps -aux|grep nodemanager –color
Warning: bad syntax, perhaps a bogus ‘-‘? See /usr/share/doc/procps-3.2.8/FAQ
hadoop 4048 0.5 15.7 2802552 158380 ? Sl 18:14 1:42 /usr/local/src/jdk1.8.0_191/bin/java -Dproc_nodemanager -Xmx1000m -Dhadoop.log.dir=/home/hadoop/hadoop-2.7.7/logs -Dyarn.log.dir=/home/hadoop/hadoop-2.7.7/logs -Dhadoop.log.file=yarn-hadoop-nodemanager-tiandong64.log -Dyarn.log.file=yarn-hadoop-nodemanager-tiandong64.log -Dyarn.home.dir= -Dyarn.id.str=hadoop -Dhadoop.root.logger=INFO,RFA -Dyarn.root.logger=INFO,RFA -Djava.library.path=/home/hadoop/hadoop-2.7.7/lib/native -Dyarn.policy.file=hadoop-policy.xml -server -Dhadoop.log.dir=/home/hadoop/hadoop-2.7.7/logs -Dyarn.log.dir=/home/hadoop/hadoop-2.7.7/logs -Dhadoop.log.file=yarn-hadoop-nodemanager-tiandong64.log -Dyarn.log.file=yarn-hadoop-nodemanager-tiandong64.log -Dyarn.home.dir=/home/hadoop/hadoop-2.7.7 -Dhadoop.home.dir=/home/hadoop/hadoop-2.7.7 -Dhadoop.root.logger=INFO,RFA -Dyarn.root.logger=INFO,RFA -Djava.library.path=/home/hadoop/hadoop-2.7.7/lib/native -classpath /home/hadoop/hadoop-2.7.7/etc/hadoop:/home/hadoop/hadoop-2.7.7/etc/hadoop:/home/hadoop/hadoop-2.7.7/etc/hadoop:/home/hadoop/hadoop-2.7.7/share/hadoop/common/lib/*:/home/hadoop/hadoop-2.7.7/share/hadoop/common/*:/home/hadoop/hadoop-2.7.7/share/hadoop/hdfs:/home/hadoop/hadoop-2.7.7/share/hadoop/hdfs/lib/*:/home/hadoop/hadoop-2.7.7/share/hadoop/hdfs/*:/home/hadoop/hadoop-2.7.7/share/hadoop/yarn/lib/*:/home/hadoop/hadoop-2.7.7/share/hadoop/yarn/*:/home/hadoop/hadoop-2.7.7/share/hadoop/mapreduce/lib/*:/home/hadoop/hadoop-2.7.7/share/hadoop/mapreduce/*:/contrib/capacity-scheduler/*.jar:/contrib/capacity-scheduler/*.jar:/home/hadoop/hadoop-2.7.7/share/hadoop/yarn/*:/home/hadoop/hadoop-2.7.7/share/hadoop/yarn/lib/*:/home/hadoop/hadoop-2.7.7/etc/hadoop/nm-config/log4j.properties org.apache.hadoop.yarn.server.nodemanager.NodeManager
注意：start-dfs 和 start-yarn.sh 这两个脚本可以使用 start-all.sh 代替
4、查看 HDFS 分布式文件系统状态
[hadoop@tiandong63 ~]$ /home/hadoop/hadoop-2.7.7/bin/hdfs dfsadmin -report
四、hadoop 的简单使用
运行 hadoop 计算任务，word count 字数统计
在 HDFS 上创建文件夹：
[hadoop@tiandong63 ~]$ /home/hadoop/hadoop-2.7.7/bin/hadoop fs -mkdir /test/input
查看文件夹：
[hadoop@tiandong63 ~]$ /home/hadoop/hadoop-2.7.7/bin/hadoop fs -ls /test/
drwxr-xr-x – hadoop supergroup 0 2018-12-30 21:40 /test/input
上传文件：
创建一个计算的文件：
[hadoop@tiandong63 ~]$ more file1.txt
welcome to beijing
my name is thunder
what is your name
上传到 HDFS 的 /test/input 文件夹中
[hadoop@tiandong63 ~]$ /home/hadoop/hadoop-2.7.7/bin/hadoop fs -put /home/hadoop/file1.txt /test/input
查看是否上传成功：
[hadoop@tiandong63 ~]$ /home/hadoop/hadoop-2.7.7/bin/hadoop fs -ls /test/input
Found 1 items
-rw-r–r– 2 hadoop supergroup 56 2018-12-30 21:40 /test/input/file1.txt
计算：
[hadoop@tiandong63 ~]$ /home/hadoop/hadoop-2.7.7/bin/hadoop jar /home/hadoop/hadoop-2.7.7/share/hadoop/mapreduce/hadoop-mapreduce-examples-2.7.7.jar wordcount /test/input /test/output
查看运行结果：
[hadoop@tiandong63 ~]$ /home/hadoop/hadoop-2.7.7/bin/hadoop fs -cat /test/output/part-r-00000
beijing 1
is 2
my 1
name 2
thunder 1
to 1
welcome 1
what 1
your 1
更多 Hadoop 相关信息见Hadoop 专题页面 https://www.linuxidc.com/topicnews.aspx?tid=13
：
