CDH5包下载:http://archive.cloudera.com/cdh5/
主机规划:
IP |
Host |
部署模块 |
进程 |
192.168.107.82 |
Hadoop-NN-01 |
NameNode ResourceManager |
NameNode DFSZKFailoverController ResourceManager |
192.168.107.83 |
Hadoop-DN-01 Zookeeper-01 |
DataNode NodeManager Zookeeper |
DataNode NodeManager JournalNode QuorumPeerMain |
192.168.107.84 |
Hadoop-DN-02 Zookeeper-02 |
DataNode NodeManager Zookeeper |
DataNode NodeManager JournalNode QuorumPeerMain |
各个进程解释:
- NameNode
- ResourceManager
- DFSZKFC:DFS Zookeeper Failover Controller 激活Standby NameNode
- DataNode
- NodeManager
- JournalNode:NameNode共享editlog结点服务(如果使用NFS共享,则该进程和所有启动相关配置接可省略)。
- QuorumPeerMain:Zookeeper主进程
目录规划:
名称 |
路径 |
$HADOOP_HOME |
/home/hadoopuser/hadoop-2.6.0-cdh5.6.0 |
Data |
$ HADOOP_HOME/data |
Log |
$ HADOOP_HOME/logs |
配置:
一、关闭防火墙(防火墙可以以后配置)
二、安装JDK(略)
三、修改HostName并配置Host(3台)
[root@Linux01 ~]# vim /etc/sysconfig/network
[root@Linux01 ~]# vim /etc/hosts 192.168.107.82 Hadoop-NN-01
192.168.107.83 Hadoop-DN-01 Zookeeper-01
192.168.107.84 Hadoop-DN-02 Zookeeper-01
四、为了安全,创建Hadoop专门登录的用户(5台)
[root@Linux01 ~]# useradd hadoopuser
[root@Linux01 ~]# passwd hadoopuser
[root@Linux01 ~]# su – hadoopuser #切换用户
五、配置SSH免密码登录(2台NameNode)
[hadoopuser@Linux05 hadoop-2.6.0-cdh5.6.0]$ ssh-keygen #生成公私钥
[hadoopuser@Linux05 hadoop-2.6.0-cdh5.6.0]$ ssh-copy-id -i ~/.ssh/id_rsa.pub hadoopuser@Hadoop-NN-01
-I 表示 input
~/.ssh/id_rsa.pub 表示哪个公钥组
或者省略为:
[hadoopuser@Linux05 hadoop-2.6.0-cdh5.6.0]$ ssh-copy-id Hadoop-NN-01(或写IP:10.10.51.231) #将公钥扔到对方服务器
[hadoopuser@Linux05 hadoop-2.6.0-cdh5.6.0]$ ssh-copy-id ”6000 Hadoop-NN-01” #如果带端口则这样写
注意修改Hadoop的配置文件 Hadoop-env.sh
export HADOOP_SSH_OPTS=”-p 6000”
[hadoopuser@Linux05 hadoop-2.6.0-cdh5.6.0]$ ssh Hadoop-NN-01 #验证(退出当前连接命令:exit、logout)
[hadoopuser@Linux05 hadoop-2.6.0-cdh5.6.0]$ ssh Hadoop-NN-01 –p 6000 #如果带端口这样写
六、配置环境变量:vi ~/.bashrc 然后 source ~/.bashrc(5台)
[hadoopuser@Linux01 ~]$ vi ~/.bashrc
# hadoop cdh5
export HADOOP_HOME=/home/hadoopuser/hadoop-2.6.0-cdh5.6.0
export PATH=$PATH:$HADOOP_HOME/sbin:$HADOOP_HOME/bin [hadoopuser@Linux01 ~]$ source ~/.bashrc #生效
七、安装zookeeper(2台DataNode)
1、解压
2、配置环境变量:vi ~/.bashrc
[hadoopuser@Linux01 ~]$ vi ~/.bashrc
# zookeeper cdh5
export ZOOKEEPER_HOME=/home/hadoopuser/zookeeper-3.4.5-cdh5.6.0
export PATH=$PATH:$ZOOKEEPER_HOME/bin [hadoopuser@Linux01 ~]$ source ~/.bashrc #生效
3、修改日志输出
[hadoopuser@Linux01 ~]$ vi $ZOOKEEPER_HOME/libexec/zkEnv.sh
56行: 找到如下位置修改语句:ZOO_LOG_DIR="$ZOOKEEPER_HOME/logs"
4、修改配置文件
[hadoopuser@Linux01 ~]$ vi $ZOOKEEPER_HOME/conf/zoo.cfg # zookeeper
tickTime=2000
initLimit=10
syncLimit=5
dataDir=/home/hadoopuser/zookeeper-3.4.5-cdh5.6.0/data
clientPort=2181 # cluster
server.1=Zookeeper-01:2888:3888
server.2=Zookeeper-02:2888:3888
5、设置myid
(1)Hadoop-DN -01:
mkdir $ZOOKEEPER_HOME/data
echo 1 > $ZOOKEEPER_HOME/data/myid
(2)Hadoop-DN -02:
mkdir $ZOOKEEPER_HOME/data
echo 2 > $ZOOKEEPER_HOME/data/myid
6、各结点启动:
[hadoopuser@Linux01 ~]$ zkServer.sh start
7、验证
[hadoopuser@Linux01 ~]$ jps 3051 Jps
2829 QuorumPeerMain
8、状态
[hadoopuser@Linux01 ~]$ zkServer.sh status JMX enabled by default
Using config: /home/zero/zookeeper/zookeeper-3.4.5-cdh5.0.1/bin/../conf/zoo.cfg
Mode: follower
9、附录zoo.cfg各配置项说明
属性 |
意义 |
tickTime |
时间单元,心跳和最低会话超时时间为tickTime的两倍 |
dataDir |
数据存放位置,存放内存快照和事务更新日志 |
clientPort |
客户端访问端口 |
initLimit |
配 置 Zookeeper 接受客户端(这里所说的客户端不是用户连接 Zookeeper服务器的客户端,而是 Zookeeper 服务器集群中连接到 Leader 的 Follower 服务器)初始化连接时最长能忍受多少个心跳时间间隔数。当已经超过 10 个心跳的时间(也就是 tickTime)长度后 Zookeeper 服务器还没有收到客户端的返回信息,那么表明这个客户端连接失败。总的时间长度就是 5*2000=10 秒。 |
syncLimit |
这个配置项标识 Leader 与 Follower 之间发送消息,请求和应答时间长度,最长不能超过多少个 |
server.id=host:port:port server.A=B:C:D |
集群结点列表: A :是一个数字,表示这个是第几号服务器; B :是这个服务器的 ip 地址; C :表示的是这个服务器与集群中的 Leader 服务器交换信息的端口; D :表示的是万一集群中的 Leader 服务器挂了,需要一个端口来重新进行选举,选出一个新的 Leader,而这个端口就是用来执行选举时服务器相互通信的端口。如果是伪集群的配置方式,由于 B 都是一样,所以不同的 Zookeeper 实例通信端口号不能一样,所以要给它们分配不同的端口号。 |
八、安装Hadoop,并配置(只装1台配置完成后分发给其它节点)
1、解压
2、修改配置文件
(1)修改 $HADOOP_HOME/etc/hadoop/masters
Hadoop-NN-01
(2)修改 $HADOOP_HOME/etc/hadoop/slaves
Hadoop-DN-01
Hadoop-DN-02
(3)修改 $HADOOP_HOME/etc/hadoop/vi core-site.xml
<configuration>
<property>
<name>fs.defaultFS</name>
<value>hdfs://Hadoop-NN-01:9000</value>
<description>定义HadoopMaster的URI和端口</description>
</property>
<property>
<name>io.file.buffer.size</name>
<value>131072</value>
<description>用作序列化文件处理时读写buffer的大小</description>
</property>
<property>
<name>hadoop.tmp.dir</name>
<value>/home/hadoopuser/hadoop-2.6.0-cdh5.6.0/tmp</value>
<description>临时数据存储目录设定</description>
</property>
</configuration>
(4)修改 $HADOOP_HOME/etc/hadoop/hdfs-site.xml
<configuration>
<property>
<name>dfs.namenode.name.dir</name>
<value>/home/hadoopuser/hadoop-2.6.0-cdh5.6.0/dfs/name</value>
<description> namenode 存放name table(fsimage)本地目录(需要修改)</description>
</property>
<property>
<name>dfs.datanode.data.dir</name>
<value>/home/hadoopuser/hadoop-2.6.0-cdh5.6.0/dfs/data</value>
<description>datanode存放block本地目录(需要修改)</description>
</property>
<property>
<name>dfs.replication</name>
<value>1</value>
<description>文件副本个数,默认为3</description>
</property>
<property>
<name>dfs.blocksize</name>
<value>134217728</value>
<description>块大小128M</description>
</property>
<property>
<name>dfs.permissions</name>
<value>false</value>
<description>是否对DFS中的文件进行权限控制(测试中一般用false)</description>
</property>
</configuration>
(5)修改 $HADOOP_HOME/etc/hadoop/yarn-site.xml
<configuration>
<property>
<name>yarn.resourcemanager.address</name>
<value>Hadoop-NN-01:8032</value>
</property>
<property>
<name>yarn.resourcemanager.scheduler.address</name>
<value>Hadoop-NN-01:8030</value>
</property>
<property>
<name>yarn.resourcemanager.resource-tracker.address</name>
<value>Hadoop-NN-01:8031</value>
</property>
<property>
<name>yarn.resourcemanager.admin.address</name>
<value>Hadoop-NN-01:8033</value>
</property>
<property>
<name>yarn.resourcemanager.webapp.address</name>
<value>Hadoop-NN-01:8088</value>
</property>
<property>
<name>yarn.nodemanager.aux-services</name>
<value>mapreduce_shuffle</value>
</property>
<property>
<name>yarn.nodemanager.aux-services.mapreduce.shuffle.class</name>
<value>org.apache.hadoop.mapred.ShuffleHandler</value>
</property>
</configuration>
(6)修改 $HADOOP_HOME/etc/hadoop/ mapred-site.xml
<configuration>
<property>
<name>mapreduce.framework.name</name>
<value>yarn</value>
</property>
<property>
<name>mapreduce.jobhistory.address</name>
<value>Hadoop-NN-01:10020</value>
</property>
<property>
<name>mapreduce.jobhistory.webapp.address</name>
<value>Hadoop-NN-01:19888</value>
</property>
</configuration>
(7)修改 $HADOOP_HOME/etc/hadoop/hadoop-env.sh
#--------------------Java Env------------------------------
export JAVA_HOME="/usr/java/jdk1.8.0_73"
#--------------------Hadoop Env----------------------------
#export HADOOP_PID_DIR=${HADOOP_PID_DIR}
export HADOOP_PREFIX="/home/hadoopuser/hadoop-2.6.0-cdh5.6.0"
#--------------------Hadoop Daemon Options-----------------
# export HADOOP_NAMENODE_OPTS="-Dhadoop.security.logger=${HADOOP_SECURITY_LOGGER:-INFO,RFAS} -Dhdfs.audit.logger=${HDFS_AUDIT_LOGGER:-INFO,NullAppender} $HADOOP_NAMENODE_OPTS"
# export HADOOP_DATANODE_OPTS="-Dhadoop.security.logger=ERROR,RFAS $HADOOP_DATANODE_OPTS"
#--------------------Hadoop Logs---------------------------
#export HADOOP_LOG_DIR=${HADOOP_LOG_DIR}/$USER
#--------------------SSH PORT-------------------------------
export HADOOP_SSH_OPTS="-p 6000" #如果你修改了SSH登录端口,一定要修改此配置。
(8)修改 $HADOOP_HOME/etc/hadoop/yarn-env.sh
#Yarn Daemon Options
#export YARN_RESOURCEMANAGER_OPTS
#export YARN_NODEMANAGER_OPTS
#export YARN_PROXYSERVER_OPTS
#export HADOOP_JOB_HISTORYSERVER_OPTS #Yarn Logs
export YARN_LOG_DIR="/home/hadoopuser/hadoop-2.6.0-cdh5.6.0/logs"
3、分发程序
scp -r /home/hadoopuser/hadoop-2.6.0-cdh5.6.0 hadoopuser@Hadoop-DN-01:/home/hadoopuser
scp -r /home/hadoopuser/hadoop-2.6.0-cdh5.6.0 hadoopuser@Hadoop-DN-02:/home/hadoopuser
4、格式化NameNode
[hadoopuser@Linux01 hadoop-2.6.0-cdh5.6.0]$ hadoop namenode -format
5、启动JournalNode:
[hadoopuser@Linux01 hadoop-2.6.0-cdh5.6.0]$ hadoop-daemon.sh start journalnode starting journalnode, logging to /home/hadoopuser/hadoop-2.6.0-cdh5.6.0/logs/hadoop-puppet-journalnode-BigData-03.out
验证JournalNode:
[hadoopuser@Linux01 hadoop-2.6.0-cdh5.6.0]$ jps 9076 Jps
9029 JournalNode
6、启动HDFS
集群启动法:Hadoop-NN-01: start-dfs.sh
[hadoopuser@Linux01 hadoop-2.6.0-cdh5.6.0]$ start-dfs.sh
单进程启动法:
<1>NameNode(Hadoop-NN-01,Hadoop-NN-02):hadoop-daemon.sh start namenode
<2>DataNode(Hadoop-DN-01,Hadoop-DN-02,Hadoop-DN-03):hadoop-daemon.sh start datanode
<3>JournalNode(Hadoop-DN-01,Hadoop-DN-02,Hadoop-DN-03):hadoop-daemon.sh start journalnode
7、启动Yarn
<1>集群启动
Hadoop-NN-01启动Yarn,命令所在目录:$HADOOP_HOME/sbin
[hadoopuser@Linux01 hadoop-2.6.0-cdh5.6.0]$ start-yarn.sh
<2>单进程启动
ResourceManager(Hadoop-NN-01,Hadoop-NN-02):yarn-daemon.sh start resourcemanager
DataNode(Hadoop-DN-01,Hadoop-DN-02,Hadoop-DN-03):yarn-daemon.sh start nodemanager
验证(略)