尚学堂大数据学习笔记（二） CentOS6.5 + JDK8 + hadoop2.6.5 搭建Hadoop集群

尚学堂大数据学习笔记（二） CentOS6.5 + JDK8 + hadoop2.6.5 安装配置HDFS

1. 安装CentOS6.5系统

工具vmware + CentOS-6.5-x86_64-bin-DVD1.iso镜像

1.1创建CentOS6.5系统虚拟机

1.1虚拟机创建

略过，可查找其他文章
使用的虚拟机镜像：

1.2 开启ssh服务

service sshd start

1.3 更改hostname

node1、node2、node3、node4改成相应名称

vi /etc/sysconfig/network

2 配置Host

  修改hosts文件（node1、node2、node3、node4）
  vi /etc/hosts

3. 安装配置JDK8

（所有节点：node1、node2、node3、node4）

3.1 上传jdk

[外链图片转存失败(img-dNNlPafG-1568773014968)(assets/1568687103215.png)]

3.2 解压

使用下面命令解压jdk包

tar -zxvf xxxxxx.tar.gz

解压完之后：

[外链图片转存失败(img-Gt60pHsq-1568773014969)(assets/1568687296529.png)]

3.3 将解压后的jdk包拷贝到/usr/java目录

[root@node1 ~]# mkdir /usr/java
[root@node1 ~]# mv jdk1.8.0_211/ /usr/java/


结果：
[root@node1 ~]# cd /usr/java/
[root@node1 java]# ll
total 4
drwxr-xr-x. 7 uucp 143 4096 Apr  1 20:51 jdk1.8.0_211
[root@node1 java]#
[root@node1 java]# cd jdk1.8.0_211/
[root@node1 jdk1.8.0_211]# pwd
/usr/java/jdk1.8.0_211

3.4 在所有节点上配置Java 环境

node1、node2、node3、node4

[root@node1 jdk1.8.0_211]# vi /etc/profile
[root@node1 jdk1.8.0_211]# 

# 在末尾处新增
export JAVA_HOME=/usr/java/jdk1.8.0_211
export PATH=$PATH:$JAVA_HOME/bin

# 重启
. /etc/profile

3.5 将node1发送到其他节点node2、node3、node4

先 
cd /usr/java/jdk1.8.0_211

在node2、node3、node4的/usr目录下创建java
mkdir /usr/java

拷贝jdk
[root@node1 jdk1.8.0_211]# scp -r /usr/java/jdk1.8.0_211/ node2:`pwd`
[root@node1 jdk1.8.0_211]# scp -r /usr/java/jdk1.8.0_211/ node3:`pwd`
[root@node1 jdk1.8.0_211]# scp -r /usr/java/jdk1.8.0_211/ node4:`pwd`

4.关闭防火墙

# 永久关闭防火墙
chkconfig iptables off
# 展示关闭防火墙
service iptables stop

5.配置免密SSH

# 在node1、node2、node3、node4生成密钥

[root@node1 jdk1.8.0_211]# ssh-keygen
Generating public/private rsa key pair.
Enter file in which to save the key (/root/.ssh/id_rsa): 
Enter passphrase (empty for no passphrase): 
Enter same passphrase again: 
Your identification has been saved in /root/.ssh/id_rsa.
Your public key has been saved in /root/.ssh/id_rsa.pub.
The key fingerprint is:
62:12:00:8f:dd:36:56:1e:e8:b8:24:07:c5:d8:cc:62 root@node1
The key's randomart image is:
+--[ RSA 2048]----+
|.Oo  .o          |
|+E=o.o .         |
|oooo* .          |
|. +o.o           |
| + .. o S        |
|  .  o .         |
|                 |
|                 |
|                 |
+-----------------+
[root@node1 jdk1.8.0_211]# 

# 处理密钥

[root@node1 jdk1.8.0_211]# cd ~
[root@node1 ~]# cd .ssh/
[root@node1 .ssh]# ll
total 12
-rw-------. 1 root root 1675 Sep 16 19:39 id_rsa
-rw-r--r--. 1 root root  392 Sep 16 19:39 id_rsa.pub
-rw-r--r--. 1 root root  397 Sep 16 12:49 known_hosts
[root@node1 .ssh]# cat id_rsa.pub >> authorized_keys
[root@node1 .ssh]# ll
total 16
-rw-r--r--. 1 root root  392 Sep 16 19:42 authorized_keys
-rw-------. 1 root root 1675 Sep 16 19:39 id_rsa
-rw-r--r--. 1 root root  392 Sep 16 19:39 id_rsa.pub
-rw-r--r--. 1 root root  397 Sep 16 12:49 known_hosts
[root@node1 .ssh]# cat authorized_keys 
ssh-rsa AAAAB3NzaC1yc2EAAAABIwAAAQEAy6IWsRhhBySt64m29Ezk0qJXpa5knI/xvw2R6rXwcfxA3sXHQYDZ4bUFCEgofQe99Kw5iCN0aztvUm6v/wSbn/5eR6Fu/gjVcC4siYOhGrKkhNLLzIkrVfba1qYjEzGmpZdA9mRMsNxqpZ7/8D3y5qXuIhqgOooggUiB7EcVjIfIUUL2k8XDHPI8CJwyNskjm+vtjxqP3f73hZBFuS4ozPAQLEM9gXQHW6kAXJn8AB2ukxxnvs1spEdHgtsFURl0U45BjjIm5Di7eUhxLJ6+E06k62XGWQcfbvIpiEYeol0FGPaE0H/3KhUwvoDM+wU6gvRu1J0T5PkWgJasBPAy8w== root@node1
[root@node1 .ssh]# 


# 配置免密
[root@node1 .ssh]# scp ./id_rsa.pub root@node2:`pwd`/node1.pub
[root@node1 .ssh]# scp ./id_rsa.pub root@node3:`pwd`/node1.pub
[root@node1 .ssh]# scp ./id_rsa.pub root@node4:`pwd`/node1.pub
[root@node2 .ssh]# cat node1.pub >> authorized_keys
[root@node3 .ssh]# cat node1.pub >> authorized_keys
[root@node4 .ssh]# cat node1.pub >> authorized_keys

[root@node2 .ssh]# scp ./id_rsa.pub root@node1:`pwd`/node2.pub
[root@node2 .ssh]# scp ./id_rsa.pub root@node3:`pwd`/node2.pub
[root@node2 .ssh]# scp ./id_rsa.pub root@node4:`pwd`/node2.pub
[root@node1 .ssh]# cat node2.pub >> authorized_keys
[root@node3 .ssh]# cat node2.pub >> authorized_keys
[root@node4 .ssh]# cat node2.pub >> authorized_keys

[root@node3 .ssh]# scp ./id_rsa.pub root@node1:`pwd`/node3.pub
[root@node3 .ssh]# scp ./id_rsa.pub root@node2:`pwd`/node3.pub
[root@node3 .ssh]# scp ./id_rsa.pub root@node4:`pwd`/node3.pub
[root@node1 .ssh]# cat node3.pub >> authorized_keys
[root@node2 .ssh]# cat node3.pub >> authorized_keys
[root@node4 .ssh]# cat node3.pub >> authorized_key

[root@node4 .ssh]# scp ./id_rsa.pub root@node1:`pwd`/node4.pub
[root@node4 .ssh]# scp ./id_rsa.pub root@node2:`pwd`/node4.pub
[root@node4 .ssh]# scp ./id_rsa.pub root@node3:`pwd`/node4.pub
[root@node1 .ssh]# cat node4.pub >> authorized_keys
[root@node2 .ssh]# cat node4.pub >> authorized_keys
[root@node3 .ssh]# cat node4.pub >> authorized_key

6. 设置集群之间时间同步

该部分转自https://blog.csdn.net/know9163/article/details/81141203

集群时间同步：在集群中找一台机器（node1,这里的node1就是任意一台机器，也可以写对应的IP地址)，然后集群中的其他机器与node1 每十分钟同步一次。

步骤：

1. rpm -qa | grep ntp 查看ntp 和ntpdate 是否安装

[root@node1 share]# rpm -qa | grep ntp
fontpackages-filesystem-1.41-1.1.el6.noarch
ntpdate-4.2.4p8-3.el6.centos.x86_64
ntp-4.2.4p8-3.el6.centos.x86_64

2. vi /etc/ntp.conf 需要修改三处

​ a 打开一个注释 ，192.168.1.0 是node1机器上的网关。

# Hosts on local network are less restricted.
restrict 192.168.1.0 mask 255.255.255.0 nomodify notrap

改成自己的网段192.168.x.0

​ b 将server0 ，server1 ，server2 ，server3 注释掉

# Use public servers from the pool.ntp.org project.
# Please consider joining the pool (http://www.pool.ntp.org/join.html).
#server 0.centos.pool.ntp.org
#server 1.centos.pool.ntp.org
#server 2.centos.pool.ntp.org
#server 3.centos.pool.ntp.org

​ c 打开两个注释，server 和 fudge

# Undisciplined Local Clock. This is a fake driver intended for backup
# and when no outside source of synchronized time is available.
server  127.127.1.0     # local clock
fudge   127.127.1.0 stratum 10

3. vi /etc/sysconfig/ntpd 加上SYNC_HWCLOCK=yes

# Drop root to id 'ntp:ntp' by default.
SYNC_HWCLOCK=yes
OPTIONS="-u ntp:ntp -p /var/run/ntpd.pid -g"

4. chkconfig ntpd on 将ntp 永久开启

5. service ntpd start 后，可以查看状态 service ntpd status

6. crontab -e 编写定时器同步时间， 意义：每十分钟与node1 同步一次时间。需要在集群中其他的机器中都编写 crontab -e

## sync cluster time
## 分 时 日 月 周 这里是每十分钟同步

0-59/10 * * * * /usr/sbin/ntpdate node1

7. ntpdate node1 然后就可以手动先同步一下时间.

7. 配置HDFS

7.0 在Windowd上配置单机Hadoop

//… 省略

7.1 配置伪分布式HDFS

在单节点上配置

1. 上传Hadoop文件

2. 解压文件

cd到上传文件目录 cd /opt/sxt
使用 tar -zxvf hadoop-2.6.5.tar.gz 解压

tar -zxvf xxxx.tar.gz 解压

[root@node1 ~]# ll
total 369816
-rw-------. 1 root root      3320 Sep 16 07:26 anaconda-ks.cfg
drwxr-xr-x. 2 root root      4096 Sep 16 07:32 Desktop
drwxr-xr-x. 2 root root      4096 Sep 16 07:32 Documents
drwxr-xr-x. 2 root root      4096 Sep 16 07:32 Downloads
drwxr-xr-x. 9 root root      4096 May 24  2017 hadoop-2.6.5
-rw-r--r--. 1 root root 183594876 Sep 16 19:45 hadoop-2.6.5.tar.gz
-rw-r--r--. 1 root root     41364 Sep 16 07:25 install.log
-rw-r--r--. 1 root root      9154 Sep 16 07:23 install.log.syslog
-rw-r--r--. 1 root root 194990602 Sep 16 19:23 jdk-8u211-linux-x64.tar.gz
drwxr-xr-x. 2 root root      4096 Sep 16 07:32 Music
drwxr-xr-x. 2 root root      4096 Sep 16 07:32 Pictures
drwxr-xr-x. 2 root root      4096 Sep 16 07:32 Public
drwxr-xr-x. 2 root root      4096 Sep 16 07:32 Templates
drwxr-xr-x. 2 root root      4096 Sep 16 07:32 Videos
[root@node1 ~]# 
[root@node1 ~]# mv hadoop-2.6.5 /opt/sxt/
[root@node1 ~]# cd /opt/sxt/
[root@node1 sxt]# ll
total 4
drwxr-xr-x. 9 root root 4096 May 24  2017 hadoop-2.6.5
[root@node1 sxt]# 

3. 修改配置文件

1. 修改profile文件

vi /etc/profile

重新加载配置文件
source /etc/profile

vi /etc/profile

# 添加hadoop配置
export JAVA_HOME=/usr/java/jdk1.8.0_211
export HADOOP_HOME=/opt/sxt/hadoop-2.6.5
export PATH=$PATH:$JAVA_HOME/bin:$HADOOP_HOME/bin:$HADOOP_HOME/sbin

# 重新加载profile文件
[root@node1 hadoop-2.6.5]# . /etc/profile
[root@node1 hadoop-2.6.5]# hd
hdfs             hdfs.cmd         hdfs-config.cmd  hdfs-config.sh   hdparm 

2.修改hadoop-env.sh

先cd到解压后的hadoop目录里面
vi etc/hadoop/hadoop-env.sh

修改JAVA_HOME为自己配置的JAVA_HOME

[root@node1 hadoop-2.6.5]# vi etc/hadoop/hadoop-env.sh 

# 修改里面JAVA_HOME配置

# The java implementation to use.
export JAVA_HOME=/usr/java/jdk1.8.0_211

3. 修改core-site.xml

vi etc/hadoop/core-site.xml

vi etc/hadoop/core-site.xml

<property>
    <name>fs.defaultFS</name>
    <value>hdfs://node1:9000</value>
</property>
<property>
    <name>hadoop.tmp.dir</name>
    <value>/var/sxt/hadoop/local</value>
</property>

4. 修改hdfs-site.xml

[root@node1 hadoop-2.6.5]#  vi etc/hadoop/hdfs-site.xml

<property>
    <name>dfs.replication</name>
    <value>1</value>
</property>
<property>
    <name>dfs.namenode.secondary.http-address</name>
    <value>node1:50090</value>
</property>

replication表示副本数， 伪分布式要设置为1

5. 修改slaves

vi etc/hadoop/slaves

[root@node1 hadoop-2.6.5]# vi etc/hadoop/slaves 

node1

6. 格式化NameNode

hdfs namenode -format

看到这句证明成功

[root@node1 hadoop-2.6.5]# hdfs namenode -format

7. 启动

start-dfs.sh

输入jps能看到以下效果代表启动成功

[root@node1 current]# start-dfs.sh 
Starting namenodes on [node1]
The authenticity of host 'node1 (192.168.219.167)' can't be established.
RSA key fingerprint is 6c:5a:b4:9a:9e:e1:27:99:9c:34:66:5c:d5:93:d0:72.
Are you sure you want to continue connecting (yes/no)? yes
node1: Warning: Permanently added 'node1,192.168.219.167' (RSA) to the list of known hosts.
node1: starting namenode, logging to /opt/sxt/hadoop-2.6.5/logs/hadoop-root-namenode-node1.out
node1: starting datanode, logging to /opt/sxt/hadoop-2.6.5/logs/hadoop-root-datanode-node1.out
Starting secondary namenodes [node1]
node1: starting secondarynamenode, logging to /opt/sxt/hadoop-2.6.5/logs/hadoop-root-secondarynamenode-node1.out
[root@node1 current]# jps
3490 Jps
3220 DataNode
3143 NameNode
3327 SecondaryNameNode
[root@node1 current]# 

8. 浏览器访问测试

在本机上使用浏览器访问 http://IP:50070可以看到：

7.2 配置完全分布式HDFS

1、 配置node2、node3、node4的/etc/profile
# 添加配置
export JAVA_HOME=/usr/java/jdk1.8.0_211
export HADOOP_HOME=/opt/sxt/hadoop-2.6.5
export PATH=$PATH:$JAVA_HOME/bin:$HADOOP_HOME/bin:$HADOOP_HOME/sbin

1. 修改master上的 core-site.xml

[root@node1 hadoop-2.6.5]# vi etc/hadoop/core-site.xml

<configuration>
    <property>
        <name>fs.defaultFS</name>
        <value>hdfs://node1:9000</value>
    </property>
    <property>
        <name>hadoop.tmp.dir</name>
        <value>/var/sxt/hadoop/full</value>
    </property>
</configuration>

2. 修改master上的hdfs-site.xml

[root@node1 hadoop-2.6.5]# vi etc/hadoop/hdfs-site.xml

<configuration>
    <property>
        <name>dfs.replication</name>
        <value>2</value>
    </property>
    <property>
        <name>dfs.namenode.secondary.http-address</name>
        <value>node2:50090</value>
    </property>
</configuration>

3. 修改master上的slaves

[root@node1 hadoop-2.6.5]# vi etc/hadoop/slaves 

node2
node3
node4

4. 使用scp命令将整个hadoop项目拷贝分发给所有子节点（node2,node3,node4）的相同目录下

先cd到master的opt目录下
scp -r ./sxt/ root@node2:/opt/
scp -r ./sxt/ root@node3:/opt/
scp -r ./sxt/ root@node4:/opt/

5.启动

1. hdfs namenode -format
2. start-dfs.sh
3. jps 使用jps查看相应模块是否启动
4. 使用浏览器访问测试，masterip:50070

7.3 搭建高可用HA

1. 配置Zookeeper

node2、node3、node4

1. 上传并解压Zookeeper

解压到/opt/sxt/下
tar -zxvf zookeeper-3.4.6.tar.gz -C /opt/sxt/

2.配置zookeeper环境（三个节点都需要）

vi /etc/profile

export JAVA_HOME=/usr/java/jdk1.8.0_211
export HADOOP_HOME=/opt/sxt/hadoop-2.6.5
export ZOOKEEPER_HOME=/opt/sxt/zookeeper-3.4.6
export PATH=$PATH:$JAVA_HOME/bin:$HADOOP_HOME/bin:$HADOOP_HOME/sbin:$ZOOKEEPER_HOME/bin

source /etc/profile

3. 修改zookeeper配置文件

先cd到conf目录下，拷贝一份配置文件，重命名为zoo.cfg：cp zoo_sample.cfg zoo.cfg
修改配置文件zoo.cfg

1.修改数据地址
# example sakes.
dataDir=/var/sxt/zk

2.在末尾追加
#autopurge.purgeInterval=1
server.1=192.168.219.155:2888:3888
server.2=192.168.219.156:2888:3888
server.3=192.168.219.157:2888:3888

4. 创建先前定义的zookeeper文件目录以及myid文件

# 1. 创建目录
mkdir -p /var/sxt/zk

# 创建myid文件，写入先前配置在末为追加的server.id ,当前服务器对应的是server.1所以写入1
echo 1 > /var/sxt/zk/myid

5. 拷贝node2当前配置好的zookeeper项目到其他节点node3，node4

先cd到/opt/sxt/目录下

scp -r ./zookeeper-3.4.6/ node3:`pwd`
scp -r ./zookeeper-3.4.6/ node4:`pwd`

6. 在每个节点上各自执行4步骤

# 1. 创建目录
mkdir -p /var/sxt/zk

# 创建myid文件，写入先前配置在末为追加的server.id ,当前服务器对应的是server.1所以写入1
echo 1 > /var/sxt/zk/myid

7.启动zookeeper

启动顺序 node2，node3，node4

启动zookeeper
zkServer.sh start

查看zookeeper状态
zkServer.sh status

2.配置HDFS

1. master配置hdfs-site.xml，按照自身配置修改节点名称，公钥名称

<configuration>
    <property>
        <name>dfs.replicationname>
        <value>2value>
    property>
<property>
  <name>dfs.nameservicesname>
  <value>myclustervalue>
property>
<property>
  <name>dfs.ha.namenodes.myclustername>
  <value>nn1,nn2value>
property>
<property>
  <name>dfs.namenode.rpc-address.mycluster.nn1name>
  <value>node1:8020value>
property>
<property>
  <name>dfs.namenode.rpc-address.mycluster.nn2name>
  <value>node2:8020value>
property>
<property>
  <name>dfs.namenode.http-address.mycluster.nn1name>
  <value>node1:50070value>
property>
<property>
  <name>dfs.namenode.http-address.mycluster.nn2name>
  <value>node2:50070value>
property>
<property>
  <name>dfs.namenode.shared.edits.dirname>
  <value>qjournal://node1:8485;node2:8485;node3:8485/myclustervalue>
property>
<property>
  <name>dfs.journalnode.edits.dirname>
  <value>/var/sxt/hadoop/ha/jnvalue>
property>
<property>
  <name>dfs.client.failover.proxy.provider.myclustername>
  <value>org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvidervalue>
property>
<property>
  <name>dfs.ha.fencing.methodsname>
  <value>sshfencevalue>
property>
<property>
  <name>dfs.ha.fencing.ssh.private-key-filesname>
  <value>/root/.ssh/id_rsavalue>
property>
<property>
   <name>dfs.ha.automatic-failover.enabledname>
   <value>truevalue>
property>
configuration>

2. master修改core-site.xml

hadoop.tmp.dir必须是一个空目录

<configuration>
    <property>
        <name>fs.defaultFSname>
        <value>hdfs://myclustervalue>
    property>
    <property>
        <name>hadoop.tmp.dirname>
        <value>/var/sxt/hadoop/havalue>
    property>
    <property>
        <name>ha.zookeeper.quorumname>
        <value>node2:2181,node3:2181,node4:2181value>
     property>
configuration>

3. master分发配置文件到所有子节点

scp hdfs-site.xml core-site.xml node2:`pwd`
scp hdfs-site.xml core-site.xml node3:`pwd`
scp hdfs-site.xml core-site.xml node4:`pwd`

4. 手动启动journalnode（master，node2，node3）

hadoop-daemon.sh start journalnode

使用jps查看journalnode是否启动成功

5.格式化

1. 在master上格式化namenode

hdfs namenode -format

2. 启动namenode

hadoop-daemon.sh start namenode

3. 启动完之后使用jps查看是否启动成功

[root@node1 hadoop]# jps
4721 Jps
4521 JournalNode
4651 NameNode

4. 对node2，node3进行同步

分别在node2，node3节点执行同步命令

hdfs namenode -bootstrapStandby

可以看到下面信息，则代表成功：

······
19/01/31 15:33:32 INFO namenode.NameNode: createNameNode [-bootstrapStandby]
=====================================================
About to bootstrap Standby ID nn2 from:
           Nameservice ID: mycluster
        Other Namenode ID: nn1
  Other NN's HTTP address: http://node1:50070
  Other NN's IPC  address: node1/192.168.219.154:8020
             Namespace ID: 1181164627
            Block pool ID: BP-2019459657-192.168.219.154-1548919768243
               Cluster ID: CID-ab317192-2eb3-42a4-81bf-b2f6e15454bc
           Layout version: -60
       isUpgradeFinalized: true
=====================================================
19/01/31 15:33:33 INFO common.Storage: Storage directory /var/sxt/hadoop/ha/dfs/name has been successfully formatted.

······

6. 使用ZKFC

1. 在node4节点上启动zookeeper客户端

zkCli.sh

2. 选择任意一个节点格式化zkfc

hdfs zkfc -formatZK

可以看到：

······
19/01/31 15:39:43 INFO ha.ActiveStandbyElector: Session connected.
19/01/31 15:39:43 INFO ha.ActiveStandbyElector: Successfully created /hadoop-ha/mycluster in ZK.
19/01/31 15:39:43 INFO zookeeper.ClientCnxn: EventThread shut down
19/01/31 15:39:43 INFO zookeeper.ZooKeeper: Session: 0x368a2b711ff0001 closed

3. 在node4上使用ls /可以看到创建了一个hadoop目录

[zk: localhost:2181(CONNECTED) 8] ls /                   
[zookeeper, hadoop-ha]

6. master启动HDFS

start-dfs.sh

可以看到：

[root@node1 hadoop]# start-dfs.sh 
Starting namenodes on [node1 node2]
node1: starting namenode, logging to /opt/sxt/hadoop-2.6.5/logs/hadoop-root-namenode-node1.out
node2: starting namenode, logging to /opt/sxt/hadoop-2.6.5/logs/hadoop-root-namenode-node2.out
node2: starting datanode, logging to /opt/sxt/hadoop-2.6.5/logs/hadoop-root-datanode-node2.out
node3: starting datanode, logging to /opt/sxt/hadoop-2.6.5/logs/hadoop-root-datanode-node3.out
node4: starting datanode, logging to /opt/sxt/hadoop-2.6.5/logs/hadoop-root-datanode-node4.out
Starting journal nodes [node1 node2 node3]
node2: starting journalnode, logging to /opt/sxt/hadoop-2.6.5/logs/hadoop-root-journalnode-node2.out
node3: starting journalnode, logging to /opt/sxt/hadoop-2.6.5/logs/hadoop-root-journalnode-node3.out
node1: starting journalnode, logging to /opt/sxt/hadoop-2.6.5/logs/hadoop-root-journalnode-node1.out
Starting ZK Failover Controllers on NN hosts [node1 node2]
node1: starting zkfc, logging to /opt/sxt/hadoop-2.6.5/logs/hadoop-root-zkfc-node1.out
node2: starting zkfc, logging to /opt/sxt/hadoop-2.6.5/logs/hadoop-root-zkfc-node2.out

8. 搭建Yarn

1. 修改master的mapred-site.xml

先拷贝一份，重命名为mapred-site.xml

 cp mapred-site.xml.template mapred-site.xml

修改mapred-site.xml配置

<configuration>
    <property>
        <name>mapreduce.framework.namename>
        <value>yarnvalue>
    property>
configuration>

2.修改master的yarn-site.xml

vi yarn-site.xml

添加如下配置，依照自己情况修改：

<configuration>


    <property>
        <name>yarn.nodemanager.aux-servicesname>
        <value>mapreduce_shufflevalue>
    property>
    <property>
        <name>yarn.resourcemanager.ha.enabledname>
        <value>truevalue>
    property>
    <property>
        <name>yarn.resourcemanager.cluster-idname>
        <value>cluster1value>
    property>
    <property>
        <name>yarn.resourcemanager.ha.rm-idsname>
        <value>rm1,rm2value>
    property>
    <property>
        <name>yarn.resourcemanager.hostname.rm1name>
        <value>node3value>
    property>
    <property>
        <name>yarn.resourcemanager.hostname.rm2name>
        <value>node4value>
    property>
    <property>
        <name>yarn.resourcemanager.zk-addressname>
        <value>node2:2181,node3:2181,node4:2181value>
    property>
configuration>

3. 将master配置好的文件分发给其他节点

scp mapred-site.xml yarn-site.xml node2:`pwd`
scp mapred-site.xml yarn-site.xml node3:`pwd`
scp mapred-site.xml yarn-site.xml node4:`pwd`

4. 启动Yarn

start-yarn.sh

可以看到：

starting yarn daemons
starting resourcemanager, logging to /opt/sxt/hadoop-2.6.5/logs/yarn-root-resourcemanager-node1.out
node4: starting nodemanager, logging to /opt/sxt/hadoop-2.6.5/logs/yarn-root-nodemanager-node4.out
node3: starting nodemanager, logging to /opt/sxt/hadoop-2.6.5/logs/yarn-root-nodemanager-node3.out
node2: starting nodemanager, logging to /opt/sxt/hadoop-2.6.5/logs/yarn-root-nodemanager-node2.out

当前命令只能启动nodemanager的节点，若是要启动node3，node4则需要在相应节点上手动启动。
分别在node3，node4执行以下命令

yarn-daemon.sh start resourcemanager

使用jps查看ResourceManager，NodeManager是否启动成功

[root@node3 ~]# jps
4434 NodeManager
3714 JournalNode
3635 DataNode
4793 Jps
4570 ResourceManager
3231 QuorumPeerMain

5. 测试

使用本机浏览器访问node3:8088 或 node4:8088
可以看到：
在这里插入图片描述