2 Hadoop伪分布模式部署 2018-05-16

Hadoop部署方式:
1.单机模式standalone： 1个java进程
2.伪分布模式Pseudo-Distributed Mode： 开发|学习 多个java进程
3.集群模式Cluster Mode ：生产 多台机器多个java进程

伪分布式部署: HDFS

1.创建hadoop服务的一个用户

[root@hadoop003 software]# useradd hadoop
[root@hadoop003 software]# id hadoop
uid=515(hadoop) gid=515(hadoop) groups=515(hadoop)
配置hadoop用户权限
[root@hadoop003 software]# vi /etc/sudoers
hadoop ALL=(root) NOPASSWD:ALL

2.部署JAVA

Oracle jdk1.8(Open JDK尽量不要使用)
2.1 解压+环境变量
上节已操作过
2.2 CDH课 /usr/java
查看Java服务路径
[root@hadoop003 jdk1.8.0_45]# which java
/usr/java/jdk1.8.0_45/bin/java

3.部署ssh服务是运行

系统安装时默认安装ssh服务
[root@hadoop003 ~]# service sshd status
openssh-daemon (pid 1386) is running...

4.解压hadoop，并进行软连接

[root@hadoop003 software]# rz hadoop-2.8.1.tar.gz上传压缩包
[root@hadoop003 software]# tar -xzvf hadoop-2.8.1.tar.gz
[root@hadoop003 software]# ln -s /opt/software/hadoop-2.8.1 hadoop
修改权限
chown -R hadoop:hadoop hadoop
chown -R hadoop:hadoop hadoop/*
chown -R hadoop:hadoop /opt/software/hadoop-2.8.1
删除无用文件
[root@hadoop003 software]# cd hadoop
[root@hadoop003 hadoop]# rm -f .txt
注意：
chown -R hadoop:hadoop 文件夹 -->文件夹和文件夹的里面的
chown -R hadoop:hadoop 软连接文件夹 --> 只修改软连接文件夹，不会修改文件夹里面的
chown -R hadoop:hadoop 软连接文件夹/ --> 软连接文件夹不修改，只会修改文件夹里面的
chown -R hadoop:hadoop hadoop-2.8.1 --> 修改原文件夹

bin/: 命令
etc/：配置文件
sbin/: 用来启动关闭hadoop进程

5.切换hadoop用户和配置

[root@hadoop003 hadoop]# su - hadoop
[hadoop@hadoop003 ~]$ ll
total 0
[hadoop@hadoop003 ~]$ cd /opt/software/hadoop
[hadoop@hadoop003 hadoop]$ ll
total 28
drwxr-xr-x. 2 hadoop hadoop 4096 Dec 10 11:54 bin
drwxr-xr-x. 3 hadoop hadoop 4096 Dec 10 11:54 etc
drwxr-xr-x. 2 hadoop hadoop 4096 Dec 10 11:54 include
drwxr-xr-x. 3 hadoop hadoop 4096 Dec 10 11:54 lib
drwxr-xr-x. 2 hadoop hadoop 4096 Dec 10 11:54 libexec
drwxr-xr-x. 2 hadoop hadoop 4096 Dec 10 11:54 sbin
drwxr-xr-x. 3 hadoop hadoop 4096 Dec 10 11:54 share
[hadoop@rzdatahadoop002 hadoop]$ cd etc/hadoop
说明：
hadoop-env.sh : hadoop配置环境
core-site.xml : hadoop 核心配置文件
hdfs-site.xml : hdfs服务的 --> 会起进程
[mapred-site.xml : mapred计算所需要的配置文件] 只当在jar计算时才有
yarn-site.xml : yarn服务的 --> 会起进程
slaves: 集群的机器名称

[hadoop@hadoop003 hadoop]$ vi core-site.xml
添加：

    
            fs.defaultFS
            hdfs://localhost:9000
    

[hadoop@hadoop003 hadoop]$ vi hdfs-site.xml
添加：

    
        dfs.replication
        1
    

6.配置hadoop用户的ssh的信任关系

[hadoop@hadoop003 ~]$ ssh-keygen -t rsa -P '' -f ~/.ssh/id_rsa
[hadoop@hadoop003 ~]$ cat ~/.ssh/id_rsa.pub >> ~/.ssh/authorized_keys
[hadoop@hadoop003 ~]$ chmod 0600 ~/.ssh/authorized_keys

第一次输入需要yes
[hadoop@hadoop003 .ssh]$ ssh hadoop003 date
The authenticity of host 'hadoop003 (192.168.137.201)' can't be established.
RSA key fingerprint is 9a:ea:f5:06:bf:de:ca:82:66:51:81:fe:bf:8a:62:36.
Are you sure you want to continue connecting (yes/no)? yes
Warning: Permanently added 'hadoop003,192.168.137.201' (RSA) to the list of known hosts.
Wed Dec 13 22:15:01 CST 2017
[hadoop@hadoop003 .ssh]$ ll
total 16
-rw-------. 1 hadoop hadoop 404 Dec 13 22:13 authorized_keys
-rw-------. 1 hadoop hadoop 1675 Dec 13 22:13 id_rsa
-rw-r--r--. 1 hadoop hadoop 404 Dec 13 22:13 id_rsa.pub
-rw-r--r--. 1 hadoop hadoop 413 Dec 13 22:15 known_hosts
第二次输入，不需要
[hadoop@hadoop003 .ssh]$ ssh hadoop003 date
Wed Dec 13 22:15:07 CST 2017

7.格式化

[hadoop@hadoop003 hadoop]$ bin/hdfs namenode -format

8.启动HDFS服务

[hadoop@hadoop003 sbin]$ ./start-dfs.sh
Starting namenodes on [localhost]
The authenticity of host 'localhost (::1)' can't be established.
RSA key fingerprint is 9a:ea:f5:06:bf:de:ca:82:66:51:81:fe:bf:8a:62:36.
Are you sure you want to continue connecting (yes/no)? yes
localhost: Warning: Permanently added 'localhost' (RSA) to the list of known hosts.
localhost: Error: JAVA_HOME is not set and could not be found.
localhost: Error: JAVA_HOME is not set and could not be found.
Starting secondary namenodes [0.0.0.0]
The authenticity of host '0.0.0.0 (0.0.0.0)' can't be established.
RSA key fingerprint is 9a:ea:f5:06:bf:de:ca:82:66:51:81:fe:bf:8a:62:36.
Are you sure you want to continue connecting (yes/no)? yes
0.0.0.0: Warning: Permanently added '0.0.0.0' (RSA) to the list of known hosts.
0.0.0.0: Error: JAVA_HOME is not set and could not be found.
[hadoop@hadoop003 sbin]$ ps -ef|grep hadoop
root 11292 11085 0 21:59 pts/1 00:00:00 su - hadoop
hadoop 11293 11292 0 21:59 pts/1 00:00:00 -bash
hadoop 11822 11293 0 22:34 pts/1 00:00:00 ps -ef
hadoop 11823 11293 0 22:34 pts/1 00:00:00 grep hadoop
[hadoop@hadoop003 sbin]$ echo $JAVA_HOME
/usr/java/jdk1.8.0_45
发现JAVA_HOME变量是存在的，无法启动HDFS服务
配置/etc/hadoop/hadoop-env.sh
[hadoop@hadoop003 sbin]$ vi ../etc/hadoop/hadoop-env.sh

The java implementation to use.添加：

export JAVA_HOME=/usr/java/jdk1.8.0_45

[hadoop@hadoop003 sbin]$ ./start-dfs.sh
Starting namenodes on [localhost]
localhost: starting namenode, logging to /opt/software/hadoop-2.8.1/logs/hadoop-hadoop-namenode-hadoop003.out
localhost: starting datanode, logging to /opt/software/hadoop-2.8.1/logs/hadoop-hadoop-datanode-hadoop003.out
Starting secondary namenodes [0.0.0.0]
0.0.0.0: starting secondarynamenode, logging to /opt/software/hadoop-2.8.1/logs/hadoop-hadoop-secondarynamenode-hadoop003.out

namenode(名称节点) : localhost
datanode(数据节点) : localhost
secondary namenode(第二名称节点): 0.0.0.0

http://localhost:50070/
默认的端口:50070
web: localhost:9000

9.使用命令(hadoop、hdfs)

[hadoop@hadoop003 bin]$ ./hdfs dfs -mkdir /user
[hadoop@hadoop003 bin]$ ./hdfs dfs -mkdir /user/hadoop

[hadoop@hadoop003 bin]$ echo "000000" > lizhigang.log
[hadoop@hadoop003 bin]$ ./hadoop fs -put lizhigang.log hdfs://localhost:9000/
[hadoop@hadoop003 bin]$
[hadoop@hadoop003 bin]$ ./hadoop fs -ls hdfs://localhost:9000/
Found 2 items
-rw-r--r-- 1 hadoop supergroup 7 2017-12-13 22:56 hdfs://localhost:9000/rz.log
drwxr-xr-x - hadoop supergroup 0 2017-12-13 22:55 hdfs://localhost:9000/user

[hadoop@hadoop003 bin]$ ./hadoop fs -ls /
Found 2 items
-rw-r--r-- 1 hadoop supergroup 7 2017-12-13 22:56 hdfs://localhost:9000/rz.log
drwxr-xr-x - hadoop supergroup 0 2017-12-13 22:55 hdfs://localhost:9000/user

10.修改hdfs://localhost:9000为hdfs://192.168.137.200:9000

[hadoop@hadoop003 bin]$ ../sbin/stop-dfs.sh 停止服务

[hadoop@hadoop003 bin]$ vi ../etc/hadoop/core-site.xml
修改为：

    
            fs.defaultFS
            hdfs://192.168.137.200:9000
    

[hadoop@hadoop003 bin]$ ./hdfs namenode -format
[hadoop@hadoop003 bin]$ ../sbin/start-dfs.sh
Starting namenodes on [hadoop003]
hadoop003: starting namenode, logging to /opt/software/hadoop-2.8.1/logs/hadoop-hadoop-namenode-hadoop003.out
localhost: starting datanode, logging to /opt/software/hadoop-2.8.1/logs/hadoop-hadoop-datanode-hadoop003.out
Starting secondary namenodes [0.0.0.0]
0.0.0.0: starting secondarynamenode, logging to /opt/software/hadoop-2.8.1/logs/hadoop-hadoop-secondarynamenode-hadoop003.out
[hadoop@hadoop003 bin]$ netstat -nlp|grep 9000
(Not all processes could be identified, non-owned process info
will not be shown, you would have to be root to see it all.)
tcp 0 0 192.168.137.200:9000 0.0.0.0:* LISTEN
14974/java

11.修改HDFS的服务以hadoop003启动

针对于datanode修改为hadoop003:
[hadoop@hadoop003 hadoop]$ vi slaves
修改为：
hadoop003

针对于secondarynamenode修改:
[hadoop@hadoop003 hadoop]$ vi hdfs-site.xml
添加：

        
                 dfs.namenode.secondary.http-address
                 hadoop003:50090
        
        
                 dfs.namenode.secondary.https-address
                 hadoop003:50091
        

[hadoop@hadoop003 hadoop]$ cd ../../sbin
[hadoop@hadoop003 sbin]$ ./stop-dfs.sh
[hadoop@hadoop003 sbin]$ ./start-dfs.sh
Starting namenodes on [hadoop003]
hadoop003: starting namenode, logging to /opt/software/hadoop-2.8.1/logs/hadoop-hadoop-namenode-hadoop003.out
hadoop003: starting datanode, logging to /opt/software/hadoop-2.8.1/logs/hadoop-hadoop-datanode-hadoop003.out
Starting secondary namenodes [hadoop003]
hadoop003: starting secondarynamenode, logging to /opt/software/hadoop-2.8.1/logs/hadoop-hadoop-secondarynamenode-hadoop003.out

12.Yarn部署

进入目录：/opt/software/hadoop/etc/hadoop
[hadoop@hadoop003 hadoop]# cp mapred-site.xml.template mapred-site.xml
[hadoop@hadoop003 hadoop]# vi mapred-site.xml
添加：

    
        mapreduce.framework.name
        yarn
    

[hadoop@hadoop003 hadoop]# vi yarn-site.xml
添加：

    
        yarn.nodemanager.aux-services
        mapreduce_shuffle
    

修改mapred-site.xml文件权限：
[hadoop@hadoop003 hadoop]# chown hadoop:hadoop *
启动服务：
[hadoop@hadoop003 hadoop]$ sbin/start-yarn.sh
web界面查看：
http://192.168.137.200:8088/
http://localhost:8088/

13.MR Job测试

查看Mapreduce 测试例程路径：
[hadoop@hadoop003 hadoop]$ find ./ -name "example"
执行：
[hadoop@hadoop003 hadoop]$ bin/hadoop jar ./share/hadoop/mapreduce/hadoop-mapreduce-examples-2.8.1.jar pi 5 10
注意：执行时Hadoop多次格式化出现17/12/19 00:19:51 WARN hdfs.DataStreamer: DataStreamer Excep
参考：https://blog.csdn.net/yu0_zhang0/article/details/78841623解决

停止Yarn：
[hadoop@hadoop003 sbin]$ ./stop-yarn.sh