GitChat 作者:鸣宇淳 原文:史上最详细的Hadoop环境搭建 关注公众号:GitChat 技术杂谈,一本正经的讲技术 【不要错过文末活动哦】
前言Hadoop在大数据技术体系中的地位至关重要,Hadoop是大数据技术的基础,对Hadoop基础知识的掌握的扎实程度,会决定在大数据技术道路上走多远。
这是一篇入门文章,Hadoop的学习方法很多,网上也有很多学习路线图。本文的思路是:以安装部署Apache Hadoop2.x版本为主线,来介绍Hadoop2.x的架构组成、各模块协同工作原理、技术细节。安装不是目的,通过安装认识Hadoop才是目的。
本文分为五个部分、十三节、四十九步。
第一部分:Linux环境安装Hadoop是运行在Linux,虽然借助工具也可以运行在Windows上,但是建议还是运行在Linux系统上,第一部分介绍Linux环境的安装、配置、Java JDK安装等。
第二部分:Hadoop本地模式安装Hadoop本地模式只是用于本地开发调试,或者快速安装体验Hadoop,这部分做简单的介绍。
第三部分:Hadoop伪分布式模式安装学习Hadoop一般是在伪分布式模式下进行。这种模式是在一台机器上各个进程上运行Hadoop的各个模块,伪分布式的意思是虽然各个模块是在各个进程上分开运行的,但是只是运行在一个操作系统上的,并不是真正的分布式。
第四部分:完全分布式安装完全分布式模式才是生产环境采用的模式,Hadoop运行在服务器集群上,生产环境一般都会做HA,以实现高可用。
第五部分:Hadoop HA安装HA是指高可用,为了解决Hadoop单点故障问题,生产环境一般都做HA部署。这部分介绍了如何配置Hadoop2.x的高可用,并简单介绍了HA的工作原理。 安装过程中,会穿插简单介绍涉及到的知识。希望能对大家有所帮助。
第一部分:Linux环境安装 第一步、配置Vmware NAT网络一、Vmware网络模式介绍
参考:http://blog.csdn.net/collection4u/article/details/14127671
二、NAT模式配置
NAT是网络地址转换,是在宿主机和虚拟机之间增加一个地址转换服务,负责外部和虚拟机之间的通讯转接和IP转换。
我们部署Hadoop集群,这里选择NAT模式,各个虚拟机通过NAT使用宿主机的IP来访问外网。
我们的要求是集群中的各个虚拟机有固定的IP、可以访问外网,所以进行如下设置:
1、 Vmware安装后,默认的NAT设置如下:
2、 默认的设置是启动DHCP服务的,NAT会自动给虚拟机分配IP,但是我们需要将各个机器的IP固定下来,所以要取消这个默认设置。
3、 为机器设置一个子网网段,默认是192.168.136网段,我们这里设置为100网段,将来各个虚拟机Ip就为 192.168.100.*。
4、 点击NAT设置按钮,打开对话框,可以修改网关地址和DNS地址。这里我们为NAT指定DNS地址。
5、 网关地址为当前网段里的.2地址,好像是固定的,我们不做修改,先记住网关地址就好了,后面会用到。
第二步、安装Linux操作系统三、Vmware上安装Linux系统
1、 文件菜单选择新建虚拟机
2、 选择经典类型安装,下一步。
3、 选择稍后安装操作系统,下一步。
4、 选择Linux系统,版本选择CentOS 64位。
5、 命名虚拟机,给虚拟机起个名字,将来显示在Vmware左侧。并选择Linux系统保存在宿主机的哪个目录下,应该一个虚拟机保存在一个目录下,不能多个虚拟机使用一个目录。
6、 指定磁盘容量,是指定分给Linux虚拟机多大的硬盘,默认20G就可以,下一步。
7、 点击自定义硬件,可以查看、修改虚拟机的硬件配置,这里我们不做修改。
8、 点击完成后,就创建了一个虚拟机,但是此时的虚拟机还是一个空壳,没有操作系统,接下来安装操作系统。
9、 点击编辑虚拟机设置,找到DVD,指定操作系统ISO文件所在位置。
10、 点击开启此虚拟机,选择第一个回车开始安装操作系统。
11、 设置root密码。
12、 选择Desktop,这样就会装一个Xwindow。
13、 先不添加普通用户,其他用默认的,就把Linux安装完毕了。
四、设置网络
因为Vmware的NAT设置中关闭了DHCP自动分配IP功能,所以Linux还没有IP,需要我们设置网络各个参数。
1、 用root进入Xwindow,右击右上角的网络连接图标,选择修改连接。
2、 网络连接里列出了当前Linux里所有的网卡,这里只有一个网卡System eth0,点击编辑。
3、 配置IP、子网掩码、网关(和NAT设置的一样)、DNS等参数,因为NAT里设置网段为100.*,所以这台机器可以设置为192.168.100.10网关和NAT一致,为192.168.100.2
4、 用ping来检查是否可以连接外网,如下图,已经连接成功。
五、修改Hostname
1、 临时修改hostname
[root@localhost Desktop]# hostname bigdata-senior01.chybinmy.com
这种修改方式,系统重启后就会失效。
2、 永久修改hostname
想永久修改,应该修改配置文件 /etc/sysconfig/network。
命令:[root@bigdata -senior01 ~] vim /etc/sysconfig/network
打开文件后,
NETWORKING=yes #使用网络
HOSTNAME=bigdata-senior01.chybinmy.com #设置主机名
六、配置Host
命令:[root@bigdata -senior01 ~] vim /etc/hosts
添加hosts: 192.168 .100 .10 bigdata-senior01.chybinmy.com
七、关闭防火墙
学习环境可以直接把防火墙关闭掉。
(1) 用root用户登录后,执行查看防火墙状态。
[root@bigdata -senior01 hadoop]# service iptables status
(2) 用[root@bigdata-senior01 hadoop]# service iptables stop关闭防火墙,这个是临时关闭防火墙。
[root@bigdata -senior01 hadoop-2.5 .0 ]# service iptables stop
iptables: Setting chains to policy ACCEPT: filter [ OK ]
iptables: Flushing firewall rules: [ OK ]
iptables: Unloading modules: [ OK ]
(3) 如果要永久关闭防火墙用。
[root@bigdata -senior01 hadoop]# chkconfig iptables off
关闭,这种需要重启才能生效。
八、关闭selinux
selinux是Linux一个子安全机制,学习环境可以将它禁用。
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ vim /etc/sysconfig/selinux
# This file controls the state of SELinux on the system.
SELINUX= can take one of these three values:
enforcing - SELinux security policy is enforced.
permissive - SELinux prints warnings instead of enforcing.
disabled - No SELinux policy is loaded.
SELINUX=disabled
SELINUXTYPE= can take one of these two values:
targeted - Targeted processes are protected ,
mls - Multi Level Security protection.
SELINUXTYPE=targeted
第三步、安装JDK九、安装Java JDK
1、 查看是否已经安装了java JDK。
[root@bigdata -senior01 Desktop]# java –version
注意:Hadoop机器上的JDK,最好是Oracle的Java JDK,不然会有一些问题,比如可能没有JPS命令。 如果安装了其他版本的JDK,卸载掉。
2、 安装java JDK
(1) 去下载Oracle版本Java JDK:jdk-7u67-linux-x64.tar.gz
(2) 将jdk-7u67-linux-x64.tar.gz解压到/opt/modules目录下
[root@bigdata -senior01 /]# tar -zxvf jdk-7 u67-linux-x64.tar.gz -C /opt/modules
(3) 添加环境变量
设置JDK的环境变量 JAVA_HOME。需要修改配置文件/etc/profile,追加
export JAVA_HOME="/opt/modules/jdk1.7.0_67"
export PATH=$JAVA_HOME/bin:$PATH
修改完毕后,执行 source /etc/profile
(4)安装后再次执行 java –version,可以看见已经安装完成。
[root@bigdata -senior01 /]# java -version
java version "1.7.0_67"
Java(TM) SE Runtime Environment (build 1.7 .0 _67-b01)
Java HotSpot(TM) 64 -Bit Server VM (build 24.65 -b04, mixed mode)
第二部分:Hadoop本地模式安装 第四步、Hadoop部署模式Hadoop部署模式有:本地模式、伪分布模式、完全分布式模式、HA完全分布式模式。
区分的依据是NameNode、DataNode、ResourceManager、NodeManager等模块运行在几个JVM进程、几个机器。
模式名称 各个模块占用的JVM进程数 各个模块运行在几个机器数上 本地模式 1个 1个 伪分布式模式 N个 1个 完全分布式模式 N个 N个 HA完全分布式 N个 N个
第五步、本地模式部署十、本地模式介绍
本地模式是最简单的模式,所有模块都运行与一个JVM进程中,使用的本地文件系统,而不是HDFS,本地模式主要是用于本地开发过程中的运行调试用。下载hadoop安装包后不用任何设置,默认的就是本地模式。
十一、解压hadoop后就是直接可以使用
1、 创建一个存放本地模式hadoop的目录
[hadoop@bigdata -senior01 modules]$ mkdir /opt/modules/hadoopstandalone</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >2 、 解压hadoop文件</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 modules]$ tar -zxf /opt/sofeware/hadoop-2.5 .0 .tar.gz -C /opt/modules/hadoopstandalone/
3、 确保JAVA_HOME环境变量已经配置好
[hadoop@bigdata -senior01 modules]$ echo ${JAVA_HOME}
/opt/modules/jdk1.7 .0 _67
十二、运行MapReduce程序,验证
我们这里用hadoop自带的wordcount例子来在本地模式下测试跑mapreduce。
1、 准备mapreduce输入文件wc.input
[hadoop@bigdata -senior01 modules]$ cat /opt/data/wc.input
hadoop mapreduce hive
hbase spark storm
sqoop hadoop hive
spark hadoop
2、 运行hadoop自带的mapreduce Demo
[hadoop@bigdata -senior01 hadoopstandalone]$ bin/hadoop jar share/hadoop/mapreduce/hadoop-mapreduce-examples-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>.jar wordcount /opt/data/wc.input output2</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/7492ce20-5cba-11e7-86d9-f17e4b747fa0" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >这里可以看到job ID中有local字样,说明是运行在本地模式下的。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >3 、 查看输出文件</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >本地模式下,mapreduce的输出是输出到本地。</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoopstandalone]$ ll output2
total 4
-rw-r--r-- 1 hadoop hadoop 60 Jul 7 12 :50 part-r-00000
-rw-r--r-- 1 hadoop hadoop 0 Jul 7 12 :50 _SUCCESS
输出目录中有_SUCCESS文件说明JOB运行成功,part-r-00000是输出结果文件。
第三部分:Hadoop伪分布式模式安装 第六步、伪分布式Hadoop部署过程十三、Hadoop所用的用户设置
1、 创建一个名字为hadoop的普通用户
[root@bigdata -senior01 ~]# useradd hadoop
[root@bigdata -senior01 ~]# passwd hadoop
2、 给hadoop用户sudo权限
[root@bigdata -senior01 ~]# vim /etc/sudoers
设置权限,学习环境可以将hadoop用户的权限设置的大一些,但是生产环境一定要注意普通用户的权限限制。
root ALL=(ALL) ALL
hadoop ALL=(root) NOPASSWD:ALL
注意:如果root用户无权修改sudoers文件,先手动为root用户添加写权限。
[root@bigdata -senior01 ~]# chmod u+w /etc/sudoers
3、 切换到hadoop用户
[root@bigdata -senior01 ~]# su - hadoop
[hadoop@bigdata -senior01 ~]$</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >4 、 创建存放hadoop文件的目录</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 ~]$ sudo mkdir /opt/modules
5、 将hadoop文件夹的所有者指定为hadoop用户
如果存放hadoop的目录的所有者不是hadoop,之后hadoop运行中可能会有权限问题,那么就讲所有者改为hadoop。
[hadoop@bigdata -senior01 ~]# sudo chown -R hadoop:hadoop /opt/modules
十四、解压Hadoop目录文件
1、 复制hadoop-2.5.0.tar.gz到/opt/modules目录下。
2、 解压hadoop-2.5.0.tar.gz
[hadoop@bigdata -senior01 ~]# cd /opt/modules
[hadoop@bigdata -senior01 hadoop]# tar -zxvf hadoop-2.5 .0 .tar.gz
十五、配置Hadoop
1、 配置Hadoop环境变量
[hadoop@bigdata -senior01 hadoop]# vim /etc/profile
追加配置:
export HADOOP_HOME="/opt/modules/hadoop-2.5.0"
export PATH=$HADOOP_HOME/bin:$HADOOP_HOME/sbin:$PATH</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >执行:source /etc/profile 使得配置生效</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >验证HADOOP_HOME参数:</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 /]$ echo $HADOOP_HOME
/opt/modules/hadoop-2.5 .0
2、 配置 hadoop-env.sh、mapred-env.sh、yarn-env.sh文件的JAVA_HOME参数
[hadoop@bigdata -senior01 ~]$ sudo vim ${HADOOP_HOME}/etc/hadoop/hadoop-env.sh
修改JAVA_HOME参数为:
export JAVA_HOME="/opt/modules/jdk1.7.0_67"
3、 配置core-site.xml
[hadoop@bigdata-senior01 ~] {HADOOP_HOME}/etc/hadoop/core-site.xml
(1) fs.defaultFS参数配置的是HDFS的地址。
< property >
< name > fs.defaultFS</ name >
< value > hdfs://bigdata-senior01.chybinmy.com:8020</ value >
</ property >
(2)hadoop.tmp.dir
配置的是Hadoop临时目录,比如HDFS的NameNode数据默认都存放这个目录下,查看*-default.xml
等默认配置文件,就可以看到很多依赖${hadoop.tmp.dir}</code>的配置。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;">默认的<code style="box-sizing: border-box; font-family: "Source Code Pro", monospace; padding: 2px 4px; font-size: 13.5px; background-color: rgba(128, 128, 128, 0.075); white-space: nowrap; border-radius: 0px;">hadoop.tmp.dir</code>是<code style="box-sizing: border-box; font-family: "Source Code Pro", monospace; padding: 2px 4px; font-size: 13.5px; background-color: rgba(128, 128, 128, 0.075); white-space: nowrap; border-radius: 0px;">/tmp/hadoop-${user.name}
,此时有个问题就是NameNode会将HDFS的元数据存储在这个/tmp目录下,如果操作系统重启了,系统会清空/tmp目录下的东西,导致NameNode元数据丢失,是个非常严重的问题,所有我们应该修改这个路径。
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ sudo mkdir -p /opt/data/tmp</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><ul style="box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><li style="box-sizing: border-box;" >将临时目录的所有者修改为hadoop</li></ul><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ sudo chown –R hadoop:hadoop /opt/data/tmp
< property >
< name > hadoop.tmp.dir</ name >
< value > /opt/data/tmp</ value >
</ property >
十六、配置、格式化、启动HDFS
1、 配置hdfs-site.xml
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ vim ${HADOOP_HOME}/etc/hadoop/hdfs-site.xml
< property >
< name > dfs.replication</ name >
< value > 1</ value >
</ property >
dfs.replication配置的是HDFS存储时的备份数量,因为这里是伪分布式环境只有一个节点,所以这里设置为1。
2、 格式化HDFS
[hadoop@bigdata -senior01 ~]$ hdfs namenode –format</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >格式化是对HDFS这个分布式文件系统中的DataNode进行分块,统计所有分块后的初始元数据的存储在NameNode中。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >格式化后,查看core-site.xml里hadoop.tmp.dir(本例是/opt/data目录)指定的目录下是否有了dfs目录,如果有,说明格式化成功。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><span style="box-sizing: border-box;" >注意:</span></p><ol style="box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><li style="box-sizing: border-box;" ><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;" >格式化时,这里注意hadoop.tmp.dir目录的权限问题,应该hadoop普通用户有读写权限才行,可以将/opt/data的所有者改为hadoop。 <br style="box-sizing: border-box;" >[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ sudo chown -R hadoop:hadoop /opt/data
查看NameNode格式化后的目录。
[hadoop@bigdata -senior01 ~]$ ll /opt/data/tmp/dfs/name/current
fsimage是NameNode元数据在内存满了后,持久化保存到的文件。
fsimage*.md5
是校验文件,用于校验fsimage的完整性。
seen_txid
是hadoop的版本
vession文件里保存:
#Mon Jul 04 17 :25 :50 CST 2016
namespaceID=2101579007
clusterID=CID-205277e6 -493 b-4601 -8e33 -c09d1d23ece4
cTime=0
storageType=NAME_NODE
blockpoolID=BP-1641019026 -127.0 .0 .1 -1467624350057
layoutVersion=-57
3、 启动NameNode
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ ${HADOOP_HOME}/sbin/hadoop-daemon.sh start namenode
starting namenode, logging to /opt/modules/hadoop-2.5 .0 /logs/hadoop-hadoop-namenode-bigdata-senior01.chybinmy.com.out
4、 启动DataNode
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ ${HADOOP_HOME}/sbin/hadoop-daemon.sh start datanode
starting datanode, logging to /opt/modules/hadoop-2.5 .0 /logs/hadoop-hadoop-datanode-bigdata-senior01.chybinmy.com.out
5、 启动SecondaryNameNode
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ ${HADOOP_HOME}/sbin/hadoop-daemon.sh start secondarynamenode
starting secondarynamenode, logging to /opt/modules/hadoop-2.5 .0 /logs/hadoop-hadoop-secondarynamenode-bigdata-senior01.chybinmy.com.out
6、 JPS命令查看是否已经启动成功,有结果就是启动成功了。
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ jps
3034 NameNode
3233 Jps
3193 SecondaryNameNode
3110 DataNode
7、 HDFS上测试创建目录、上传、下载文件
HDFS上创建目录
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ ${HADOOP_HOME}/bin/hdfs dfs -mkdir /demo1
上传本地文件到HDFS上
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ ${HADOOP_HOME}/bin/hdfs dfs -put
${HADOOP_HOME}/etc/hadoop/core-site.xml /demo1</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >读取HDFS上的文件内容</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ ${HADOOP_HOME}/bin/hdfs dfs -cat /demo1/core-site.xml</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/fe96fd30-5cba-11e7-8185-21ba04c77532" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >从HDFS上下载文件到本地</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ bin/hdfs dfs -get /demo1/core-site.xml
十七、配置、启动YARN1、 配置mapred-site.xml
默认没有mapred-site.xml文件,但是有个mapred-site.xml.template配置模板文件。复制模板生成mapred-site.xml。
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]# cp etc/hadoop/mapred-site.xml.template etc/hadoop/mapred-site.xml
添加配置如下:
< property >
< name > mapreduce.framework.name</ name >
< value > yarn</ value >
</ property >
指定mapreduce运行在yarn框架上。
2、 配置yarn-site.xml
添加配置如下:
< property >
< name > yarn.nodemanager.aux-services</ name >
< value > mapreduce_shuffle</ value >
</ property >
< property >
< name > yarn.resourcemanager.hostname</ name >
< value > bigdata-senior01.chybinmy.com</ value >
</ property >
3、 启动Resourcemanager
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ ${HADOOP_HOME}/sbin/yarn-daemon.sh start resourcemanager
4、 启动nodemanager
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ ${HADOOP_HOME}/sbin/yarn-daemon.sh start nodemanager
5、 查看是否启动成功
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ jps
3034 NameNode
4439 NodeManager
4197 ResourceManager
4543 Jps
3193 SecondaryNameNode
3110 DataNode
可以看到ResourceManager、NodeManager已经启动成功了。
6、 YARN的Web页面
YARN的Web客户端端口号是8088,通过http://192.168.100.10:8088/ 可以查看。
十八、运行MapReduce Job
在Hadoop的share目录里,自带了一些jar包,里面带有一些mapreduce实例小例子,位置在share/hadoop/mapreduce/hadoop-mapreduce-examples-2.5.0.jar,可以运行这些例子体验刚搭建好的Hadoop平台,我们这里来运行最经典的WordCount实例。
1、 创建测试用的Input文件
创建输入目录:
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ bin/hdfs dfs -mkdir -p /wordcountdemo/input</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >创建原始文件:</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >在本地/opt/data目录创建一个文件wc.input,内容如下。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/178763e0-5cbe-11e7-86d9-f17e4b747fa0" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >将wc.input文件上传到HDFS的/wordcountdemo/input目录中:</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ bin/hdfs dfs -put /opt/data/wc.input /wordcountdemo/input
2、 运行WordCount MapReduce Job
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ bin/yarn jar share/hadoop/mapreduce/hadoop-mapreduce-examples-
2.5 .0 .jar wordcount /wordcountdemo/input /wordcountdemo/output
3、 查看输出结果目录
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ bin/hdfs dfs -ls /wordcountdemo/output
-rw-r--r-- 1 hadoop supergroup 0 2016 -07 -05 05 :12 /wordcountdemo/output/_SUCCESS
-rw-r--r-- 1 hadoop supergroup 60 2016 -07 -05 05 :12 /wordcountdemo/output/part-r-00000
output目录中有两个文件,_SUCCESS文件是空文件,有这个文件说明Job执行成功。
part-r-00000文件是结果文件,其中-r-说明这个文件是Reduce阶段产生的结果,mapreduce程序执行时,可以没有reduce阶段,但是肯定会有map阶段,如果没有reduce阶段这个地方有是-m-。
一个reduce会产生一个part-r-开头的文件。
查看输出文件内容。
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ bin/hdfs dfs -cat /wordcountdemo/output/part-r-00000
hadoop 3
hbase 1
hive 2
mapreduce 1
spark 2
sqoop 1
storm 1
结果是按照键值排好序的。
十九、停止Hadoop
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh stop namenode
stopping namenode
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh stop datanode
stopping datanode
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ sbin/yarn-daemon.sh stop resourcemanager
stopping resourcemanager
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ sbin/yarn-daemon.sh stop nodemanager
stopping nodemanager
二十、 Hadoop各个功能模块的理解
1、 HDFS模块
HDFS负责大数据的存储,通过将大文件分块后进行分布式存储方式,突破了服务器硬盘大小的限制,解决了单台机器无法存储大文件的问题,HDFS是个相对独立的模块,可以为YARN提供服务,也可以为HBase等其他模块提供服务。
2、 YARN模块
YARN是一个通用的资源协同和任务调度框架,是为了解决Hadoop1.x中MapReduce里NameNode负载太大和其他问题而创建的一个框架。
YARN是个通用框架,不止可以运行MapReduce,还可以运行Spark、Storm等其他计算框架。
3、 MapReduce模块
MapReduce是一个计算框架,它给出了一种数据处理的方式,即通过Map阶段、Reduce阶段来分布式地流式处理数据。它只适用于大数据的离线处理,对实时性要求很高的应用不适用。
第七步、开启历史服务二十一、历史服务介绍
Hadoop开启历史服务可以在web页面上查看Yarn上执行job情况的详细信息。可以通过历史服务器查看已经运行完的Mapreduce作业记录,比如用了多少个Map、用了多少个Reduce、作业提交时间、作业启动时间、作业完成时间等信息。
二十二、开启历史服务
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ sbin/mr-jobhistory-daemon.sh start historyserver、</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >开启后,可以通过Web页面查看历史服务器:</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><a target="_blank" href="http://bigdata-senior01.chybinmy.com:19888/" style="color: rgb(79, 161, 219); text-decoration-line: none; box-sizing: border-box;" >http:
2.5 .0 .jar wordcount /wordcountdemo/input /wordcountdemo/output1
2、 job执行中
3、 查看job历史
历史服务器的Web端口默认是19888,可以查看Web界面。
但是在上面所显示的某一个Job任务页面的最下面,Map和Reduce个数的链接上,点击进入Map的详细信息页面,再查看某一个Map或者Reduce的详细日志是看不到的,是因为没有开启日志聚集服务。
二十四、开启日志聚集
4、 日志聚集介绍
MapReduce是在各个机器上运行的,在运行过程中产生的日志存在于各个机器上,为了能够统一查看各个机器的运行日志,将日志集中存放在HDFS上,这个过程就是日志聚集。
5、 开启日志聚集
配置日志聚集功能:
Hadoop默认是不启用日志聚集的。在yarn-site.xml文件里配置启用日志聚集。
< property >
< name > yarn.log-aggregation-enable</ name >
< value > true</ value >
</ property >
< property >
< name > yarn.log-aggregation.retain-seconds</ name >
< value > 106800</ value >
</ property >
yarn.log-aggregation-enable:是否启用日志聚集功能。
yarn.log-aggregation.retain-seconds:设置日志保留时间,单位是秒。
将配置文件分发到其他节点:
[hadoop@bigdata -senior01 hadoop]$ scp /opt/modules/hadoop-2.5 .0 /etc/hadoop/yarn-site.xml bigdata-senior02.chybinmy.com:/opt/modules/hadoop-2.5 .0 /etc/hadoop/
[hadoop@bigdata -senior01 hadoop]$ scp /opt/modules/hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>/etc/hadoop/yarn-site.xml bigdata-senior03.chybinmy.com:/opt/modules/hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>/etc/hadoop/</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >重启Yarn进程:</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ sbin/stop-yarn.sh
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ sbin/start-yarn.sh</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >重启HistoryServer进程:</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ sbin/mr-jobhistory-daemon.sh stop historyserver
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ sbin/mr-jobhistory-daemon.sh start historyserver</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >6 、 测试日志聚集</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >运行一个demo MapReduce,使之产生日志:</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >bin/yarn jar share/hadoop/mapreduce/hadoop-mapreduce-examples-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>.jar wordcount /input /output1</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >查看日志:</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >运行Job后,就可以在历史服务器Web页面查看各个Map和Reduce的日志了。 <br style="box-sizing: border-box;" ></p><h3 id="第四部分完全分布式安装-1" style="margin: 0.8em 0px; padding: 0px; box-sizing: border-box; font-family: "microsoft yahei"; font-weight: 300; line-height: 1.1; color: rgb(63, 63, 63); font-size: 1.7em;" ><a name="t4" ></a><a target="_blank" name="t17" style="color: rgb(79, 161, 219); box-sizing: border-box;" ></a>第四部分:完全分布式安装</h3><h4 id="第八步完全布式环境部署hadoop" style="margin: 0.8em 0px; padding: 0px; box-sizing: border-box; font-family: "microsoft yahei"; font-weight: 300; line-height: 1.1; color: rgb(63, 63, 63); font-size: 1.25em;" ><a target="_blank" name="t18" style="color: rgb(79, 161, 219); box-sizing: border-box;" ></a>第八步、完全布式环境部署Hadoop</h4><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >完全分部式是真正利用多台Linux主机来进行部署Hadoop,对Linux机器集群进行规划,使得Hadoop各个模块分别部署在不同的多台机器上。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >二十五、环境准备</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >1 、 克隆虚拟机</p><ul style="box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><li style="box-sizing: border-box;" ><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;" >Vmware左侧选中要克隆的机器,这里对原有的BigData01机器进行克隆,虚拟机菜单中,选中管理菜单下的克隆命令。</p></li><li style="box-sizing: border-box;" ><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;" >选择“创建完整克隆”,虚拟机名称为BigData02,选择虚拟机文件保存路径,进行克隆。</p></li><li style="box-sizing: border-box;" ><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;" >再次克隆一个名为BigData03的虚拟机。</p></li></ul><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >2 、 配置网络</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >修改网卡名称:</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >在BigData02和BigData03机器上编辑网卡信息。执行sudo vim /etc/udev/rules.d/70 -persistent-net.rules命令。因为是从BigData01机器克隆来的,所以会保留BigData01的网卡eth0,并且再添加一个网卡eth1。并且eth0的Mac地址和BigData01的地址是一样的,Mac地址不允许相同,所以要删除eth0,只保留eth1网卡,并且要将eth1改名为eth0。将修改后的eth0的mac地址复制下来,修改network-scripts文件中的HWADDR属性。</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >sudo vim /etc/sysconfig/network-scripts/ifcfg-eth0</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/692fc750-5cbe-11e7-8ca5-edc6aa6f5290" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >修改网络参数:</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >BigData02机器IP改为192.168 .100 .12 </p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >BigData03机器IP改为192.168 .100 .13 </p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >3 、 配置Hostname</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >BigData02配置hostname为 bigdata-senior02.chybinmy.com</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >BigData03配置hostname为 bigdata-senior03.chybinmy.com</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >4 、 配置hosts</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >BigData01、BigData02、BigData03三台机器hosts都配置为:</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-xml hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ sudo vim /etc/hosts
192.168 .100 .10 bigdata-senior01.chybinmy.com
192.168 .100 .12 bigdata-senior02.chybinmy.com
192.168 .100 .13 bigdata-senior03.chybinmy.com
5、 配置Windows上的SSH客户端
在本地Windows中的SSH客户端上添加对BigData02、BigData03机器的SSH链接。
二十六、服务器功能规划
bigdata-senior01.chybinmy.com bigdata-senior02.chybinmy.com bigdata-senior03.chybinmy.com NameNode ResourceManage DataNode DataNode DataNode NodeManager NodeManager NodeManager HistoryServer SecondaryNameNode
二十七、在第一台机器上安装新的Hadoop
为了和之前BigData01机器上安装伪分布式Hadoop区分开来,我们将BigData01上的Hadoop服务都停止掉,然后在一个新的目录/opt/modules/app下安装另外一个Hadoop。 我们采用先在第一台机器上解压、配置Hadoop,然后再分发到其他两台机器上的方式来安装集群。
6、 解压Hadoop目录:
[hadoop@bigdata-senior01 modules]$ tar -zxf /opt/sofeware/hadoop-2.5.0.tar.gz -C /opt/modules/app/</code > <ul class ="pre-numbering" style ="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" > <li style ="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" > 1</li > </ul > </pre > <p style ="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" > 7、 配置Hadoop JDK路径修改hadoop-env.sh、mapred-env.sh、yarn-env.sh文件中的JDK路径:</p > <pre class ="prettyprint" name ="code" style ="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" > <code class ="language-xml hljs has-numbering" style ="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" > export JAVA_HOME="/opt/modules/jdk1.7.0_67"</code > <ul class ="pre-numbering" style ="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" > <li style ="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" > 1</li > </ul > </pre > <p style ="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" > 8、 配置core-site.xml</p > <pre class ="prettyprint" name ="code" style ="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" > <code class ="language-xml hljs has-numbering" style ="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" > [hadoop@bigdata-senior01 hadoop-2.5.0]$ vim etc/hadoop/core-site.xml
< configuration >
< property >
< name > fs.defaultFS</ name >
< value > hdfs://bigdata-senior01.chybinmy.com:8020</ value >
</ property >
< property >
< name > hadoop.tmp.dir</ name >
< value > /opt/modules/app/hadoop-2.5.0/data/tmp</ value >
</ property >
</ configuration >
fs.defaultFS为NameNode的地址。
hadoop.tmp.dir为hadoop临时目录的地址,默认情况下,NameNode和DataNode的数据文件都会存在这个目录下的对应子目录下。应该保证此目录是存在的,如果不存在,先创建。
9、 配置hdfs-site.xml
[hadoop@bigdata-senior01 hadoop-2.5.0]$ vim etc/hadoop/hdfs-site.xml
< configuration >
< property >
< name > dfs.namenode.secondary.http-address</ name >
< value > bigdata-senior03.chybinmy.com:50090</ value >
</ property >
</ configuration >
dfs.namenode.secondary.http-address是指定secondaryNameNode的http访问地址和端口号,因为在规划中,我们将BigData03规划为SecondaryNameNode服务器。
所以这里设置为:bigdata-senior03.chybinmy.com:50090
10、 配置slaves
[hadoop@bigdata-senior01 hadoop-2.5.0]$ vim etc/hadoop/slaves
bigdata-senior01.chybinmy.com
bigdata-senior02.chybinmy.com
bigdata-senior03.chybinmy.com
slaves文件是指定HDFS上有哪些DataNode节点。
11、 配置yarn-site.xml
[hadoop@bigdata-senior01 hadoop-2.5.0]$ vim etc/hadoop/yarn-site.xml
< property >
< name > yarn.nodemanager.aux-services</ name >
< value > mapreduce_shuffle</ value >
</ property >
< property >
< name > yarn.resourcemanager.hostname</ name >
< value > bigdata-senior02.chybinmy.com</ value >
</ property >
< property >
< name > yarn.log-aggregation-enable</ name >
< value > true</ value >
</ property >
< property >
< name > yarn.log-aggregation.retain-seconds</ name >
< value > 106800</ value >
</ property >
根据规划yarn.resourcemanager.hostname
这个指定resourcemanager服务器指向bigdata-senior02.chybinmy.com
。
yarn.log-aggregation-enable
是配置是否启用日志聚集功能。
yarn.log-aggregation.retain-seconds
是配置聚集的日志在HDFS上最多保存多长时间。
12、 配置mapred-site.xml
从mapred-site.xml.template复制一个mapred-site.xml文件。
[hadoop@bigdata-senior01 hadoop-2.5.0]$ cp etc/hadoop/mapred-site.xml.template etc/hadoop/mapred-site.xml
< configuration >
< property >
< name > mapreduce.framework.name</ name >
< value > yarn</ value >
</ property >
< property >
< name > mapreduce.jobhistory.address</ name >
< value > bigdata-senior01.chybinmy.com:10020</ value >
</ property >
< property >
< name > mapreduce.jobhistory.webapp.address</ name >
< value > bigdata-senior01.chybinmy.com:19888</ value >
</ property >
</ configuration >
mapreduce.framework.name设置mapreduce任务运行在yarn上。
mapreduce.jobhistory.address是设置mapreduce的历史服务器安装在BigData01机器上。
mapreduce.jobhistory.webapp.address是设置历史服务器的web页面地址和端口号。
二十八、设置SSH无密码登录
Hadoop集群中的各个机器间会相互地通过SSH访问,每次访问都输入密码是不现实的,所以要配置各个机器间的
SSH是无密码登录的。
1、 在BigData01上生成公钥
[hadoop@bigdata-senior01 hadoop-2.5.0]$ ssh-keygen -t rsa</code > <ul class ="pre-numbering" style ="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" > <li style ="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" > 1</li > </ul > </pre > <p style ="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" > 一路回车,都设置为默认值,然后再当前用户的Home目录下的<code style ="box-sizing: border-box; font-family: "Source Code Pro", monospace; padding: 2px 4px; font-size: 13.5px; background-color: rgba(128, 128, 128, 0.075); white-space: nowrap; border-radius: 0px;" > .ssh</code > 目录中会生成公钥文件<code style ="box-sizing: border-box; font-family: "Source Code Pro", monospace; padding: 2px 4px; font-size: 13.5px; background-color: rgba(128, 128, 128, 0.075); white-space: nowrap; border-radius: 0px;" > (id_rsa.pub)</code > 和私钥文件<code style ="box-sizing: border-box; font-family: "Source Code Pro", monospace; padding: 2px 4px; font-size: 13.5px; background-color: rgba(128, 128, 128, 0.075); white-space: nowrap; border-radius: 0px;" > (id_rsa)</code > 。</p > <p style ="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" > 2、 分发公钥</p > <pre class ="prettyprint" name ="code" style ="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" > <code class ="language-xml hljs has-numbering" style ="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" > [hadoop@bigdata-senior01 hadoop-2.5.0]$ ssh-copy-id bigdata-senior01.chybinmy.com
[hadoop@bigdata-senior01 hadoop-2.5.0]$ ssh-copy-id bigdata-senior02.chybinmy.com
[hadoop@bigdata-senior01 hadoop-2.5.0]$ ssh-copy-id bigdata-senior03.chybinmy.com</code > <ul class ="pre-numbering" style ="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" > <li style ="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" > 1</li > <li style ="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" > 2</li > <li style ="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" > 3</li > </ul > </pre > <p style ="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" > 3、 设置BigData02、BigData03到其他机器的无密钥登录</p > <p style ="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" > 同样的在BigData02、BigData03上生成公钥和私钥后,将公钥分发到三台机器上。</p > <p style ="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" > 二十九、分发Hadoop文件</p > <p style ="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" > 1、 首先在其他两台机器上创建存放Hadoop的目录</p > <pre class ="prettyprint" name ="code" style ="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" > <code class ="language-xml hljs has-numbering" style ="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" > [hadoop@bigdata-senior02 ~]$ mkdir /opt/modules/app
[hadoop@bigdata-senior03 ~]$ mkdir /opt/modules/app</code > <ul class ="pre-numbering" style ="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" > <li style ="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" > 1</li > <li style ="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" > 2</li > </ul > </pre > <p style ="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" > 2、 通过Scp分发</p > <p style ="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" > Hadoop根目录下的share/doc目录是存放的hadoop的文档,文件相当大,建议在分发之前将这个目录删除掉,可以节省硬盘空间并能提高分发的速度。</p > <p style ="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" > doc目录大小有1.6G。</p > <pre class ="prettyprint" name ="code" style ="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" > <code class ="language-xml hljs has-numbering" style ="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" > [hadoop@bigdata-senior01 hadoop-2.5.0]$ du -sh /opt/modules/app/hadoop-2.5.0/share/doc
1.6G /opt/modules/app/hadoop-2.5.0/share/doc
[hadoop@bigdata-senior01 hadoop-2.5.0]$ scp -r /opt/modules/app/hadoop-2.5.0/ bigdata-senior02.chybinmy.com:/opt/modules/app
[hadoop@bigdata-senior01 hadoop-2.5.0]$ scp -r /opt/modules/app/hadoop-2.5.0/ bigdata-senior03.chybinmy.com:/opt/modules/app</code > <ul class ="pre-numbering" style ="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" > <li style ="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" > 1</li > <li style ="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" > 2</li > <li style ="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" > 3</li > <li style ="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" > 4</li > </ul > </pre > <p style ="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" > 三十、格式NameNode</p > <p style ="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" > 在NameNode机器上执行格式化:</p > <pre class ="prettyprint" name ="code" style ="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" > <code class ="language-xml hljs has-numbering" style ="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" > [hadoop@bigdata-senior01 hadoop-2.5.0]$ /opt/modules/app/hadoop-2.5.0/bin/hdfs namenode –format
注意:
如果需要重新格式化NameNode,需要先将原来NameNode和DataNode下的文件全部删除,不然会报错,NameNode和DataNode所在目录是在core-site.xml
中hadoop.tmp.dir
、dfs.namenode.name.dir
、dfs.datanode.data.dir
属性配置的。
< property >
< name > hadoop.tmp.dir</ name >
< value > /opt/data/tmp</ value >
</ property >
< property >
< name > dfs.namenode.name.dir</ name >
< value > file://${hadoop.tmp.dir}/dfs/name</ value >
</ property >
< property >
< name > dfs.datanode.data.dir</ name >
< value > file://${hadoop.tmp.dir}/dfs/data</ value >
</ property >
因为每次格式化,默认是创建一个集群ID,并写入NameNode和DataNode的VERSION文件中(VERSION文件所在目录为dfs/name/current 和 dfs/data/current),重新格式化时,默认会生成一个新的集群ID,如果不删除原来的目录,会导致namenode中的VERSION文件中是新的集群ID,而DataNode中是旧的集群ID,不一致时会报错。
另一种方法是格式化时指定集群ID参数,指定为旧的集群ID。
三十一、启动集群
1、 启动HDFS
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ /opt/modules/app/hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>/sbin/start-dfs.sh</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/7c4619c0-5cbe-11e7-8ca5-edc6aa6f5290" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >2 、 启动YARN</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ /opt/modules/app/hadoop-2.5 .0 /sbin/start-yarn.sh
在BigData02上启动ResourceManager:
[hadoop@bigdata -senior02 hadoop-2.5 .0 ]$ sbin/yarn-daemon.sh start resourcemanager</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/98a56fd0-5cbe-11e7-8ca5-edc6aa6f5290" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >3 、 启动日志服务器</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >因为我们规划的是在BigData03服务器上运行MapReduce日志服务,所以要在BigData03上启动。</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior03 ~]$ /opt/modules/app/hadoop-2.5 .0 /sbin/mr-jobhistory-daemon.sh start historyserver
starting historyserver, logging to /opt/modules/app/hadoop-2.5 .0 /logs/mapred-hadoop-historyserver-bigda ta-senior03.chybinmy.com.out
[hadoop@bigdata -senior03 ~]$ jps
3570 Jps
3537 JobHistoryServer
3310 SecondaryNameNode
3213 DataNode
3392 NodeManager
4、 查看HDFS Web页面
http://bigdata-senior01.chybinmy.com:50070/
5、 查看YARN Web 页面
http://bigdata-senior02.chybinmy.com:8088/cluster
三十二、测试Job
我们这里用hadoop自带的wordcount例子来在本地模式下测试跑mapreduce。
1、 准备mapreduce输入文件wc.input
[hadoop@bigdata -senior01 modules]$ cat /opt/data/wc.input
hadoop mapreduce hive
hbase spark storm
sqoop hadoop hive
spark hadoop
2、 在HDFS创建输入目录input
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ bin/hdfs dfs -mkdir /input</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >3 、 将wc.input上传到HDFS</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ bin/hdfs dfs -put /opt/data/wc.input /input/wc.input
4、 运行hadoop自带的mapreduce Demo
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ bin/yarn jar share/hadoop/mapreduce/hadoop-mapreduce-examples-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>.jar wordcount /input/wc.input /output</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/a5cbfbc0-5cbe-11e7-8185-21ba04c77532" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >5 、 查看输出文件</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ bin/hdfs dfs -ls /output
Found 2 items
-rw-r--r-- 3 hadoop supergroup 0 2016 -07 -14 16 :36 /output/_SUCCESS
-rw-r--r-- 3 hadoop supergroup 60 2016 -07 -14 16 :36 /output/part-r-00000
第五部分:Hadoop HA安装HA的意思是High Availability高可用,指当当前工作中的机器宕机后,会自动处理这个异常,并将工作无缝地转移到其他备用机器上去,以来保证服务的高可用。
HA方式安装部署才是最常见的生产环境上的安装部署方式。Hadoop HA是Hadoop 2.x中新添加的特性,包括NameNode HA 和 ResourceManager HA。因为DataNode和NodeManager本身就是被设计为高可用的,所以不用对他们进行特殊的高可用处理。
第九步、时间服务器搭建
Hadoop对集群中各个机器的时间同步要求比较高,要求各个机器的系统时间不能相差太多,不然会造成很多问题。可以配置集群中各个机器和互联网的时间服务器进行时间同步,但是在实际生产环境中,集群中大部分服务器是不能连接外网的,这时候可以在内网搭建一个自己的时间服务器(NTP服务器),集群的各个机器与这个时间服务器进行时间同步。
三十三、配置NTP服务器
我们选择第三台机器(bigdata-senior03.chybinmy.com)为NTF服务器,其他机器和这台机器进行同步。
1、 检查ntp服务是否已经安装
[hadoop@bigdata -senior03 data]$ sudo rpm -qa | grep ntp
ntpdate-4.2 .6 p5-1. el6.centos.x86_64
ntp-4.2 .6 p5-1. el6.centos.x86_64
显示已经安装过了ntp程序,其中ntpdate-4.2.6p5-1.el6.centos.x86_64
是用来和某台服务器进行同步的,ntp-4.2.6p5-1.el6.centos.x86_64
是用来提供时间同步服务的。
2、 修改配置文件ntp.conf
[hadoop@bigdata -senior03 data]$ vim /etc/ntp.conf
启用restrice,修改网段
restrict 192.168.100.0 mask 255.255.255.0 nomodify notrap 将这行的注释去掉,并且将网段改为集群的网段,我们这里是100网段。
注释掉server域名配置
#server 0. centos.pool.ntp.org iburst
server 1. centos.pool.ntp.org iburst
server 2. centos.pool.ntp.org iburst
server 3. centos.pool.ntp.org iburst 是时间服务器的域名,这里不需要连接互联网,所以将他们注释掉。
修改
server 127.127.1.0
fudge 127.127.1.0 stratum 10
3、 修改配置文件ntpd
[hadoop@bigdata -senior03 ~]$ sudo vim /etc/sysconfig/ntpd</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >添加一行配置:SYNC_CLOCK=yes</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/c8db0cf0-5cbe-11e7-b8ee-cfb7eda16afc" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >4 、 启动ntp服务</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior03 ~]$ sudo chkconfig ntpd on
这样每次机器启动时,ntp服务都会自动启动。
三十四、配置其他机器的同步
切换到root用户进行配置通过contab进行定时同步:
[root@bigdata -senior02 hadoop]# crontab -e
/10 * * * /usr/sbin/ntpdate bigdata-senior03.chybinmy.com
[root@bigdata -senior02 hadoop]# crontab -e
/10 * * * /usr/sbin/ntpdate bigdata-senior03.chybinmy.com
三十五、 测试同步是否有效
1、 查看目前三台机器的时间
[hadoop@bigdata -senior03 ~]$ date "+%Y-%m-%d %H:%M:%S"
2016 -09 -23 16 :43 :56
[hadoop@bigdata -senior02 ~]$ date "+%Y-%m-%d %H:%M:%S"
2016 -09 -23 16 :44 :08
[hadoop@bigdata -senior01 data]$ date "+%Y-%m-%d %H:%M:%S"
2016 -09 -23 16 :44 :18
2、 修改bigdata-senior01上的时间
将时间改为一个以前的时间:
[hadoop@bigdata -senior01 data]$ sudo date -s '2016-01-01 00:00:00'
Fri Jan 1 00 :00 :00 CST 2016
[hadoop@bigdata -senior01 data]$ date "+%Y-%m-%d %H:%M:%S"
2016 -01 -01 00 :00 :05
等10分钟,看是否可以实现自动同步,将bigdata-senior01上的时间修改为和bigdata-senior03上的一致。
3、 查看是否自动同步时间
[hadoop@bigdata -senior01 data]$ date "+%Y-%m-%d %H:%M:%S"
2016 -09 -23 16 :54 :36
可以看到bigdata-senior01上的时间已经实现自动同步了。
第十步、Zookeeper分布式机器部署三十六、zookeeper说明
Zookeeper在Hadoop集群中的作用。
Zookeeper是分布式管理协作框架,Zookeeper集群用来保证Hadoop集群的高可用,(高可用的含义是:集群中就算有一部分服务器宕机,也能保证正常地对外提供服务。)
Zookeeper保证高可用的原理。
Zookeeper集群能够保证NamaNode服务高可用的原理是:Hadoop集群中有两个NameNode服务,两个NaameNode都定时地给Zookeeper发送心跳,告诉Zookeeper我还活着,可以提供服务,单某一个时间只有一个是Action状态,另外一个是Standby状态,一旦Zookeeper检测不到Action NameNode发送来的心跳后,就切换到Standby状态的NameNode上,将它设置为Action状态,所以集群中总有一个可用的NameNode,达到了NameNode的高可用目的。
Zookeeper的选举机制。
Zookeeper集群也能保证自身的高可用,保证自身高可用的原理是,Zookeeper集群中的各个机器分为Leader和Follower两个角色,写入数据时,要先写入Leader,Leader同意写入后,再通知Follower写入。客户端读取数时,因为数据都是一样的,可以从任意一台机器上读取数据。
这里Leader角色就存在单点故障的隐患,高可用就是解决单点故障隐患的。Zookeeper从机制上解决了Leader的单点故障问题,Leader是哪一台机器是不固定的,Leader是选举出来的。选举流程是,集群中任何一台机器发现集群中没有Leader时,就推荐自己为Leader,其他机器来同意,当超过一半数的机器同意它为Leader时,选举结束,所以Zookeeper集群中的机器数据必须是奇数。这样就算当Leader机器宕机后,会很快选举出新的Leader,保证了Zookeeper集群本身的高可用。
写入高可用。
集群中的写入操作都是先通知Leader,Leader再通知Follower写入,实际上当超过一半的机器写入成功后,就认为写入成功了,所以就算有些机器宕机,写入也是成功的。
读取高可用。
zookeeperk客户端读取数据时,可以读取集群中的任何一个机器。所以部分机器的宕机并不影响读取。
zookeeper服务器必须是奇数台,因为zookeeper有选举制度,角色有:领导者、跟随者、观察者,选举的目的是保证集群中数据的一致性。
三十七、安装zookeeper
我们这里在BigData01、BigData02、BigData03三台机器上安装zookeeper集群。
1、 解压安装包
在BigData01上安装解压zookeeper安装包。
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ tar -zxf /opt/sofeware/zookeeper-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >3.4 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.8 </span>.tar.gz -C /opt/modules/</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >2 、 修改配置</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >拷贝conf下的zoo_sample.cfg副本,改名为zoo.cfg。zoo.cfg是zookeeper的配置文件:</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 zookeeper-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >3.4 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.8 </span>]$ cp conf/zoo_sample.cfg conf/zoo.cfg
dataDir属性设置zookeeper的数据文件存放的目录:
dataDir=/opt/modules/zookeeper-3.4.8/data/zData
指定zookeeper集群中各个机器的信息:
server.1 =bigdata-senior01.chybinmy.com:2888 :3888
server.2 =bigdata-senior02.chybinmy.com:2888 :3888
server.3 =bigdata-senior03.chybinmy.com:2888 :3888
server后面的数字范围是1到255,所以一个zookeeper集群最多可以有255个机器。
3、 创建myid文件
在dataDir所指定的目录下创一个名为myid的文件,文件内容为server点后面的数字。
4、 分发到其他机器
[hadoop@bigdata -senior01 zookeeper-3.4 .8 ]$ scp -r /opt/modules/zookeeper-3.4 .8 bigdata-senior02.chybinmy.com:/opt/modules
[hadoop@bigdata -senior01 zookeeper-3.4 .8 ]$ scp -r /opt/modules/zookeeper-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >3.4 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.8 </span> bigdata-senior03.chybinmy.com:/opt/modules</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >5 、 修改其他机器上的myid文件</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-shell hljs ruby has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-variable" style="color: rgb(102, 0, 102); box-sizing: border-box;" >@bigdata </span>-senior02 zookeeper-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >3.4 </span>.<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >8 </span>]<span class="hljs-variable" style="color: rgb(102, 0, 102); box-sizing: border-box;" >$ echo 2 > /opt/modules /zookeeper-3.4 .8 /data /zData/myid
[hadoop@bigdata -senior02 zookeeper-3.4 . 8 ]$ cat /opt/modules/zookeeper-3.4 . 8 /data/zData/myid
2
[hadoop@bigdata -senior03 ~]$ echo 3 > /opt/modules /zookeeper-3.4.8/data /zData/myid
[hadoop@bigdata -senior03 ~]$ cat /opt/modules/zookeeper-3.4 .8 /data/zData/myid
3
6、 启动zookeeper
需要在各个机器上分别启动zookeeper。
[hadoop@bigdata -senior01 zookeeper-3.4 .8 ]$ bin/zkServer.sh start
[hadoop@bigdata -senior02 zookeeper-3.4 .8 ]$ bin/zkServer.sh start
[hadoop@bigdata -senior03 zookeeper-3.4 .8 ]$ </span>bin/zk Server.sh start</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;"><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;">1</li ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;">3</li ></ul></pre ><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/e6e83a10-5cbe-11e7-86d9-f17e4b747fa0" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;">三十八、zookeeper命令</p ><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><span style="box-sizing: border-box;" >进入zookeeper Shell </span></p ><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >在zookeeper根目录下执行 bin/zkCli.sh进入zk shell模式。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;">zookeeper很像一个小型的文件系统,/ 是根目录,下面的所有节点都叫zNode。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;"><span style="box-sizing: border-box;">进入zk shell 后输入任意字符,可以列出所有的zookeeper命令</span ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;"><img src="http:/ /images.gitbook.cn/f 0fc55e0 -5 cbe-11 e7-b8ee-cfb7eda16afc" alt=" enter image description here" title=" " style=" border: none; box-sizing: border-box; max-width: 100 %;"></p><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;"><span style=" box-sizing: border-box;">查询zNode上的数据:get /zookeeper</span></p><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;"><span style=" box-sizing: border-box;">创建一个zNode : create /znode1 “demodata “</span></p><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;"><span style=" box-sizing: border-box;">列出所有子zNode:ls /</span></p><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;"><img src=" http: //images.gitbook.cn/f 9ff01b0 -5 cbe-11 e7-8185 -21 ba04c77532" alt=" enter image description here" title=" " style=" border: none; box-sizing: border-box; max-width: 100 %;"></p><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;"><span style=" box-sizing: border-box;">删除znode : rmr /znode1</span></p><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;"><span style=" box-sizing: border-box;">退出shell模式:quit</span></p><h4 id=" 第十一步hadoop-2 x-hdfs-ha-部署" style=" margin: 0 .8 em 0px; padding: 0px; box-sizing: border-box; font-family: "microsoft yahei"; font-weight: 300 ; line-height: 1.1 ; color: rgb(63 , 63 , 63 ); font-size: 1.25 em;"><a target=" _blank" name=" t21" style=" color: rgb(79 , 161 , 219 ); box-sizing: border-box;"></a>第十一步、Hadoop 2.x HDFS HA 部署</h4><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;">三十九、HDFS HA原理</p><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;">单NameNode的缺陷存在单点故障的问题,如果NameNode不可用,则会导致整个HDFS文件系统不可用。所以需要设计高可用的HDFS(Hadoop HA)来解决NameNode单点故障的问题。解决的方法是在HDFS集群中设置多个NameNode节点。但是一旦引入多个NameNode,就有一些问题需要解决。</p><ul style=" box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;"><li style=" box-sizing: border-box;"><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;">HDFS HA需要保证的四个问题:</p><ul style=" box-sizing: border-box;"><li style=" box-sizing: border-box;"><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;">保证NameNode内存中元数据数据一致,并保证编辑日志文件的安全性。</p></li><li style=" box-sizing: border-box;"><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;">多个NameNode如何协作</p></li><li style=" box-sizing: border-box;"><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;">客户端如何能正确地访问到可用的那个NameNode。</p></li><li style=" box-sizing: border-box;"><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;">怎么保证任意时刻只能有一个NameNode处于对外服务状态。</p></li></ul></li><li style=" box-sizing: border-box;"><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;">解决方法</p><ul style=" box-sizing: border-box;"><li style=" box-sizing: border-box;"><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;">对于保证NameNode元数据的一致性和编辑日志的安全性,采用Zookeeper来存储编辑日志文件。</p></li><li style=" box-sizing: border-box;"><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;">两个NameNode一个是Active状态的,一个是Standby状态的,一个时间点只能有一个Active状态的 <br style=" box-sizing: border-box;">NameNode提供服务,两个NameNode上存储的元数据是实时同步的,当Active的NameNode出现问题时,通过Zookeeper实时切换到Standby的NameNode上,并将Standby改为Active状态。</p></li><li style=" box-sizing: border-box;"><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;">客户端通过连接一个Zookeeper的代理来确定当时哪个NameNode处于服务状态。</p></li></ul></li></ul><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;">四十、HDFS HA架构图</p><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;"><img src=" http: //images.gitbook.cn/ 0446 fec0 -5 cbf-11 e7-86 d9-f17e4b747fa0 " alt=" enter image description here" title=" " style=" border: none; box-sizing: border-box; max-width: 100 %;"></p><ul style=" box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;"><li style=" box-sizing: border-box;"><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;">HDFS HA架构中有两台NameNode节点,一台是处于活动状态(Active)为客户端提供服务,另外一台处于热备份状态(Standby)。</p></li><li style=" box-sizing: border-box;"><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;">元数据文件有两个文件:fsimage和edits,备份元数据就是备份这两个文件。JournalNode用来实时从Active NameNode上拷贝edits文件,JournalNode有三台也是为了实现高可用。</p></li><li style=" box-sizing: border-box;"><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;">Standby NameNode不对外提供元数据的访问,它从Active NameNode上拷贝fsimage文件,从JournalNode上拷贝edits文件,然后负责合并fsimage和edits文件,相当于SecondaryNameNode的作用。最终目的是保证Standby NameNode上的元数据信息和Active NameNode上的元数据信息一致,以实现热备份。</p></li><li style=" box-sizing: border-box;"><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;">Zookeeper来保证在Active NameNode失效时及时将Standby NameNode修改为Active状态。</p></li><li style=" box-sizing: border-box;"><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;">ZKFC(失效检测控制)是Hadoop里的一个Zookeeper客户端,在每一个NameNode节点上都启动一个ZKFC进程,来监控NameNode的状态,并把NameNode的状态信息汇报给Zookeeper集群,其实就是在Zookeeper上创建了一个Znode节点,节点里保存了NameNode状态信息。当NameNode失效后,ZKFC检测到报告给Zookeeper,Zookeeper把对应的Znode删除掉,Standby ZKFC发现没有Active状态的NameNode时,就会用shell命令将自己监控的NameNode改为Active状态,并修改Znode上的数据。 <br style=" box-sizing: border-box;">Znode是个临时的节点,临时节点特征是客户端的连接断了后就会把znode删除,所以当ZKFC失效时,也会导致切换NameNode。</p></li><li style=" box-sizing: border-box;"><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box;">DataNode会将心跳信息和Block汇报信息同时发给两台NameNode,DataNode只接受Active NameNode发来的文件读写操作指令。</p></li></ul><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;">四十一、搭建HDFS HA 环境</p><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;">1、 服务器角色规划</p><table style=" box-sizing: border-box; width: 945.778 px; border-collapse: collapse; border-spacing: 0px; border: 1 px solid rgb(238 , 238 , 238 ); color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;"><thead style=" box-sizing: border-box;"><tr style=" box-sizing: border-box;"><th style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );"><span style=" box-sizing: border-box;">bigdata-senior01.chybinmy.com</span></th><th style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );"><span style=" box-sizing: border-box;">bigdata-senior01.chybinmy.com</span></th><th style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );"><span style=" box-sizing: border-box;">bigdata-senior01.chybinmy.com</span></th></tr></thead><tbody style=" box-sizing: border-box;"><tr style=" box-sizing: border-box;"><td style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );">NameNode</td><td style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );">NameNode</td><td style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );"> </td></tr><tr style=" box-sizing: border-box;"><td style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );">Zookeeper</td><td style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );">Zookeeper</td><td style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );">Zookeeper</td></tr><tr style=" box-sizing: border-box;"><td style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );">DataNode</td><td style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );">DataNode</td><td style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );">DataNode</td></tr><tr style=" box-sizing: border-box;"><td style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );"> </td><td style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );">ResourceManage</td><td style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );">ResourceManage</td></tr><tr style=" box-sizing: border-box;"><td style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );">NodeManager</td><td style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );">NodeManager</td><td style=" box-sizing: border-box; padding: 8 px; line-height: 20 px; vertical-align: top; border: 1 px solid rgb(238 , 238 , 238 );">NodeManager</td></tr></tbody></table><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;">2、 创建HDFS HA 版本Hadoop程序目录</p><p style=" margin-top: 0px; margin-bottom: 1.1 em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63 , 63 , 63 ); font-family: "microsoft yahei"; font-size: 15 px;">在bigdata01、bigdata02、bigdata03三台机器上分别创建目录/opt/modules/hadoopha/用来存放Hadoop HA环境。</p><pre class=" prettyprint" name=" code" style=" white-space: nowrap; word-wrap: break -word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1 em; font-family: "Source Code Pro ", monospace; padding: 5 px 5 px 5 px 60 px; font-size: 14 px; line-height: 1.45 ; word-break: break -all; color: rgb(51 , 51 , 51 ); background-color: rgba(128 , 128 , 128 , 0 .05 ); border: 0px solid rgb(136 , 136 , 136 ); border-radius: 0px;"><code class=" language-java hljs has-numbering" style=" display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro ", monospace;font-size: undefined; white-space: pre; border-radius: 0px; word-wrap: normal;">[hadoop<span class=" hljs-annotation" style=" color: rgb(155 , 133 , 157 ); box-sizing: border-box;">@bigdata</span>-senior01 modules]$ mkdir /opt/modules/hadoopha
3、 新解压Hadoop 2.5.0
[hadoop@bigdata -senior01 ~]$ tar -zxf /opt/sofeware/hadoop-2.5 .0 .tar.gz -C /opt/modules/hadoopha/
4、 配置Hadoop JDK路径
修改hadoop-env.sh、mapred-env.sh、yarn-env.sh文件中的JDK路径
export JAVA_HOME="/opt/modules/jdk1.7.0_67"
5、 配置hdfs-site.xml
<xml version="1.0" encoding="UTF-8">
< configuration >
< property >
< name > dfs.nameservices</ name >
< value > ns1</ value >
</ property >
< property >
< name > dfs.ha.namenodes.ns1</ name >
< value > nn1,nn2</ value >
</ property >
< property >
< name > dfs.namenode.rpc-address.ns1.nn1</ name >
< value > bigdata-senior01.chybinmy.com:8020</ value >
</ property >
< property >
< name > dfs.namenode.rpc-address.ns1.nn2</ name >
< value > bigdata-senior02.chybinmy.com:8020</ value >
</ property >
< property >
< name > dfs.namenode.http-address.ns1.nn1</ name >
< value > bigdata-senior01.chybinmy.com:50070</ value >
</ property >
< property >
< name > dfs.namenode.http-address.ns1.nn2</ name >
< value > bigdata-senior02.chybinmy.com:50070</ value >
</ property >
< property >
< name > dfs.namenode.shared.edits.dir</ name >
< value > qjournal://bigdata-senior01.chybinmy.com:8485;bigdata-senior02.chybinmy.com:8485;bigdata-senior03.chybinmy.com:8485/ns1</ value >
</ property >
< property >
< name > dfs.journalnode.edits.dir</ name >
< value > /opt/modules/hadoopha/hadoop-2.5.0/tmp/data/dfs/jn</ value >
</ property >
< property >
< name > dfs.client.failover.proxy.provider.ns1</ name >
< value > org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider</ value >
</ property >
< property >
< name > dfs.ha.fencing.methods</ name >
< value > sshfence</ value >
</ property >
< property >
< name > dfs.ha.fencing.ssh.private-key-files</ name >
< value > /home/hadoop/.ssh/id_rsa</ value >
</ property >
</ configuration >
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 6、 配置core-site.xml
<xml version="1.0" encoding="UTF-8">
< configuration >
< property >
< name > fs.defaultFS</ name >
< value > hdfs://ns1</ value >
</ property >
< property >
< name > hadoop.tmp.dir</ name >
< value > /opt/modules/hadoopha/hadoop-2.5.0/data/tmp</ value >
</ property >
</ configuration >
hadoop.tmp.dir
设置hadoop临时目录地址,默认时,NameNode和DataNode的数据存在这个路径下。
7、 配置slaves文件
bigdata-senior01.chybinmy.com
bigdata-senior02.chybinmy.com
bigdata-senior03.chybinmy.com
8、 分发到其他节点
分发之前先将share/doc目录删除,这个目录中是帮助文件,并且很大,可以删除。
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ scp -r /opt/modules/hadoopha bigdata-senior02.chybinmy.com:/opt/modules
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ scp -r /opt/modules/hadoopha bigdata-senior03.chybinmy.com:/opt/modules</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >9 、 启动HDFS HA集群</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >三台机器分别启动Journalnode。</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ sbin/hadoop-daemon.sh start journalnode
[hadoop@bigdata -senior02 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh start journalnode
[hadoop@bigdata -senior03 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh start journalnode</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >3 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >jps命令查看是否启动。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >10 、 启动Zookeeper</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >在三台节点上启动Zookeeper:</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 zookeeper-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >3.4 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.8 </span>]$ bin/zkServer.sh start
[hadoop@bigdata -senior02 zookeeper-3.4 .8 ]$ bin/zkServer.sh start
[hadoop@bigdata -senior03 zookeeper-3.4 .8 ]$ bin/zkServer.sh start</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >3 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >11 、 格式化NameNode</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >在第一台上进行NameNode格式化:</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ bin/hdfs namenode -format
在第二台NameNode上:
[hadoop@bigdata -senior02 hadoop-2.5 .0 ]$ bin/hdfs namenode -bootstrapStandby</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >12 、 启动NameNode</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >在第一台、第二台上启动NameNode:</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ sbin/hadoop-daemon.sh start namenode
[hadoop@bigdata -senior02 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh start namenode</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >查看HDFS Web页面,此时两个NameNode都是standby状态。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >切换第一台为active状态:</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ bin/hdfs haadmin -transitionToActive nn1
可以添加上forcemanual参数,强制将一个NameNode转换为Active状态。
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ bin/hdfs haadmin –transitionToActive -forcemanual nn1</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >此时从web 页面就看到第一台已经是active状态了。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/14fd6830-5cbf-11e7-b8ee-cfb7eda16afc" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >13 、 配置故障自动转移</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >利用zookeeper集群实现故障自动转移,在配置故障自动转移之前,要先关闭集群,不能在HDFS运行期间进行配置。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><span style="box-sizing: border-box;" >关闭NameNode、DataNode、JournalNode、zookeeper</span></p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ sbin/hadoop-daemon.sh stop namenode
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh stop datanode
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh stop journalnode
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ cd ../../zookeeper-3.4 .8 /
[hadoop@bigdata -senior01 zookeeper-3.4 .8 ]$ bin/zkServer.sh stop
[hadoop@bigdata -senior02 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh stop namenode
[hadoop@bigdata -senior02 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh stop datanode
[hadoop@bigdata - senior02 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh stop journalnode
[hadoop@bigdata - senior02 hadoop-2.5 .0 ]$ cd ../../zookeeper-3.4 .8 /
[hadoop@bigdata - senior02 zookeeper-3.4 .8 ]$ bin/zkServer.sh stop
[hadoop@bigdata - senior03 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh stop datanode
[hadoop@bigdata - senior03 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh stop journalnode
[hadoop@bigdata - senior03 hadoop-2.5 .0 ]$ cd ../../zookeeper-3.4 .8 /
[hadoop@bigdata - senior03 zookeeper-3.4 .8 ]$ bin/zkServer.sh stop
修改hdfs-site.xml
< property >
< name > dfs.ha.automatic-failover.enabled</ name >
< value > true</ value >
</ property >
修改core-site.xml
< property >
< name > ha.zookeeper.quorum</ name >
< value > bigdata-senior01.chybinmy.com:2181,bigdata-senior02.chybinmy.com:2181,bigdata-senior03.chybinmy.com:2181</ value >
</ property >
将hdfs-site.xml和core-site.xml分发到其他机器
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ scp /opt/modules/hadoopha/hadoop-2.5 .0 /etc/hadoop/hdfs-site.xml bigdata-senior02.chybinmy.com:/opt/modules/hadoopha/hadoop-2.5 .0 /etc/hadoop/
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ scp /opt/modules/hadoopha/hadoop-2.5 .0 /etc/hadoop/hdfs-site.xml bigdata-senior03.chybinmy.com:/opt/modules/hadoopha/hadoop-2.5 .0 /etc/hadoop/
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ scp /opt/modules/hadoopha/hadoop-2.5 .0 /etc/hadoop/core-site.xml bigdata-senior02.chybinmy.com:/opt/modules/hadoopha/hadoop-2.5 .0 /etc/hadoop/
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ scp /opt/modules/hadoopha/hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>/etc/hadoop/core-site.xml bigdata-senior03.chybinmy.com:/opt/modules/hadoopha/hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>/etc/hadoop/</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >3 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >4 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><span style="box-sizing: border-box;" >启动zookeeper</span></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >三台机器启动zookeeper</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ /opt/modules/zookeeper-3.4 .8 /bin/zkServer.sh start
创建一个zNode
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ cd /opt/modules/hadoopha/hadoop-2.5 .0 /
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ bin/hdfs zkfc -formatZK</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/26b6e9c0-5cbf-11e7-86d9-f17e4b747fa0" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >在Zookeeper上创建一个存储namenode相关的节点。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >14 、 启动HDFS、JournalNode、zkfc</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >启动NameNode、DataNode、JournalNode、zkfc</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ sbin/start-dfs.sh
zkfc只针对NameNode监听。
四十二、测试HDFS HA
1、 测试故障自动转移和数据是否共享
在nn1上上传文件
目前bigdata-senior01节点上的NameNode是Active状态的。
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ bin/hdfs dfs -put /opt/data/wc.input /</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/3f42f0b0-5cbf-11e7-b8ee-cfb7eda16afc" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><span style="box-sizing: border-box;" >将nn1上的NodeNode进程杀掉</span></p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ kill -9 3364
nn1上的namenode已经无法访问了。
查看nn2是否是Active状态
在nn2上查看是否看见文件
经以上验证,已经实现了nn1和nn2之间的文件同步和故障自动转移。
第十二步、Hadoop 2.x YARN HA 部署四十三、YARN HA原理
Hadoop2.4版本之前,ResourceManager也存在单点故障的问题,也需要实现HA来保证ResourceManger的高可也用性。
ResouceManager从记录着当前集群的资源分配情况和JOB的运行状态,YRAN HA 利用Zookeeper等共享存储介质来存储这些信息来达到高可用。另外利用Zookeeper来实现ResourceManager自动故障转移。
MasterHADaemon:控制RM的 Master的启动和停止,和RM运行在一个进程中,可以接收外部RPC命令。
共享存储:Active Master将信息写入共享存储,Standby Master读取共享存储信息以保持和Active Master同步。
ZKFailoverController:基于Zookeeper实现的切换控制器,由ActiveStandbyElector和HealthMonitor组成,ActiveStandbyElector负责与Zookeeper交互,判断所管理的Master是进入Active还是Standby;HealthMonitor负责监控Master的活动健康情况,是个监视器。
Zookeeper:核心功能是维护一把全局锁控制整个集群上只有一个Active的ResourceManager。
四十四、搭建YARN HA环境
1、 服务器角色规划
bigdata-senior01.chybinmy.com bigdata-senior01.chybinmy.com bigdata-senior01.chybinmy.com NameNode NameNode Zookeeper Zookeeper Zookeeper DataNode DataNode DataNode ResourceManage ResourceManage NodeManager NodeManager NodeManager
2、 修改配置文件yarn-site.xml
<xml version="1.0" encoding="UTF-8">
< configuration >
< property >
< name > yarn.nodemanager.aux-services</ name >
< value > mapreduce_shuffle</ value >
</ property >
< property >
< name > yarn.log-aggregation-enable</ name >
< value > true</ value >
</ property >
< property >
< name > yarn.log-aggregation.retain-seconds</ name >
< value > 106800</ value >
</ property >
< property >
< name > yarn.resourcemanager.ha.enabled</ name >
< value > true</ value >
</ property >
< property >
< name > yarn.resourcemanager.cluster-id</ name >
< value > yarn-cluster</ value >
</ property >
< property >
< name > yarn.resourcemanager.ha.rm-ids</ name >
< value > rm12,rm13</ value >
</ property >
< property >
< name > yarn.resourcemanager.hostname.rm12</ name >
< value > bigdata-senior02.chybinmy.com</ value >
</ property >
< property >
< name > yarn.resourcemanager.hostname.rm13</ name >
< value > bigdata-senior03.chybinmy.com</ value >
</ property >
< property >
< name > yarn.resourcemanager.zk-address</ name >
< value > bigdata-senior01.chybinmy.com:2181,bigdata-senior02.chybinmy.com:2181,bigdata-senior03.chybinmy.com:2181</ value >
</ property >
< property >
< name > yarn.resourcemanager.recovery.enabled</ name >
< value > true</ value >
</ property >
< property >
< name > yarn.resourcemanager.store.class</ name >
< value > org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore</ value >
</ property >
</ configuration >
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 3、 分发到其他机器
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ scp /opt/modules/hadoopha/hadoop-2.5 .0 /etc/hadoop/yarn-site.xml bigdata-senior02.chybinmy.com:/opt/modules/hadoopha/hadoop-2.5 .0 /etc/hadoop/
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ scp /opt/modules/hadoopha/hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>/etc/hadoop/yarn-site.xml bigdata-senior03.chybinmy.com:/opt/modules/hadoopha/hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>/etc/hadoop/\</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >4 、 启动</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >在bigdata-senior01上启动yarn:</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ sbin/start-yarn.sh
在bigdata-senior02、bigdata-senior03上启动resourcemanager:
[hadoop@bigdata -senior02 hadoop-2.5 .0 ]$ sbin/yarn-daemon.sh start resourcemanager
[hadoop@bigdata -senior03 hadoop-2.5 .0 ]$ sbin/yarn-daemon.sh start resourcemanager</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >启动后各个节点的进程。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/a5661890-5cbf-11e7-8185-21ba04c77532" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/a9db9b20-5cbf-11e7-8ca5-edc6aa6f5290" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/ae2049b0-5cbf-11e7-86d9-f17e4b747fa0" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >Web客户端访问bigdata02机器上的resourcemanager正常,它是active状态的。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><a target="_blank" href="http://bigdata-senior02.chybinmy.com:8088/cluster" style="color: rgb(79, 161, 219); text-decoration-line: none; box-sizing: border-box;" >http:
6、 在job运行过程中,将Active状态的resourcemanager进程杀掉。
[hadoop@bigdata -senior02 hadoop-2.5 .0 ]$ kill -<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >9 </span> <span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >4475 </span></code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >7 、 观察另外一个resourcemanager是否可以自动接替。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >bigdata02的resourcemanage Web客户端已经不能访问,bigdata03的resourcemanage已经自动变为active状态。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >8 、 观察job是否可以顺利完成。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >而mapreduce job 也能顺利完成,没有因为resourcemanager的意外故障而影响运行。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >经过以上测试,已经验证YARN HA 已经搭建成功。</p><h4 id="第十三步hdfs-federation-架构部署" style="margin: 0.8em 0px; padding: 0px; box-sizing: border-box; font-family: "microsoft yahei"; font-weight: 300; line-height: 1.1; color: rgb(63, 63, 63); font-size: 1.25em;" ><a target="_blank" name="t23" style="color: rgb(79, 161, 219); box-sizing: border-box;" ></a>第十三步、HDFS Federation 架构部署</h4><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >四十六、HDFS Federation 的使用原因</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >1 、 单个NameNode节点的局限性</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >命名空间的限制。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >NameNode上存储着整个HDFS上的文件的元数据,NameNode是部署在一台机器上的,因为单个机器硬件的限制,必然会限制NameNode所能管理的文件个数,制约了数据量的增长。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >数据隔离问题。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >整个HDFS上的文件都由一个NameNode管理,所以一个程序很有可能会影响到整个HDFS上的程序,并且权限控制比较复杂。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >性能瓶颈。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >单个NameNode时HDFS文件系统的吞吐量受限于单个NameNode的吞吐量。因为NameNode是个JVM进程,JVM进程所占用的内存很大时,性能会下降很多。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >2 、 HDFS Federation介绍</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >HDFS Federation是可以在Hadoop集群中设置多个NameNode,不同于HA中多个NameNode是完全一样的,是多个备份,Federation中的多个NameNode是不同的,可以理解为将一个NameNode切分为了多个NameNode,每一个NameNode只负责管理一部分数据。 <br style="box-sizing: border-box;" >HDFS Federation中的多个NameNode共用DataNode。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >四十七、HDFS Federation的架构图</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/cbde6c70-5cbf-11e7-8185-21ba04c77532" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >四十八、HDFS Federation搭建</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >1 、 服务器角色规划</p><table style="box-sizing: border-box; width: 945.778px; border-collapse: collapse; border-spacing: 0px; border: 1px solid rgb(238, 238, 238); color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><thead style="box-sizing: border-box;" ><tr style="box-sizing: border-box;" ><th style="box-sizing: border-box; padding: 8px; line-height: 20px; vertical-align: top; border: 1px solid rgb(238, 238, 238);" ><span style="box-sizing: border-box;" >bigdata-senior01.chybinmy.com</span></th><th style="box-sizing: border-box; padding: 8px; line-height: 20px; vertical-align: top; border: 1px solid rgb(238, 238, 238);" ><span style="box-sizing: border-box;" >bigdata-senior01.chybinmy.com</span></th><th style="box-sizing: border-box; padding: 8px; line-height: 20px; vertical-align: top; border: 1px solid rgb(238, 238, 238);" ><span style="box-sizing: border-box;" >bigdata-senior01.chybinmy.com</span></th></tr></thead><tbody style="box-sizing: border-box;" ><tr style="box-sizing: border-box;" ><td style="box-sizing: border-box; padding: 8px; line-height: 20px; vertical-align: top; border: 1px solid rgb(238, 238, 238);" >NameNode1</td><td style="box-sizing: border-box; padding: 8px; line-height: 20px; vertical-align: top; border: 1px solid rgb(238, 238, 238);" >NameNode2</td><td style="box-sizing: border-box; padding: 8px; line-height: 20px; vertical-align: top; border: 1px solid rgb(238, 238, 238);" >NameNode3</td></tr><tr style="box-sizing: border-box;" ><td style="box-sizing: border-box; padding: 8px; line-height: 20px; vertical-align: top; border: 1px solid rgb(238, 238, 238);" > </td><td style="box-sizing: border-box; padding: 8px; line-height: 20px; vertical-align: top; border: 1px solid rgb(238, 238, 238);" >ResourceManage</td><td style="box-sizing: border-box; padding: 8px; line-height: 20px; vertical-align: top; border: 1px solid rgb(238, 238, 238);" > </td></tr><tr style="box-sizing: border-box;" ><td style="box-sizing: border-box; padding: 8px; line-height: 20px; vertical-align: top; border: 1px solid rgb(238, 238, 238);" >DataNode</td><td style="box-sizing: border-box; padding: 8px; line-height: 20px; vertical-align: top; border: 1px solid rgb(238, 238, 238);" >DataNode</td><td style="box-sizing: border-box; padding: 8px; line-height: 20px; vertical-align: top; border: 1px solid rgb(238, 238, 238);" >DataNode</td></tr><tr style="box-sizing: border-box;" ><td style="box-sizing: border-box; padding: 8px; line-height: 20px; vertical-align: top; border: 1px solid rgb(238, 238, 238);" >NodeManager</td><td style="box-sizing: border-box; padding: 8px; line-height: 20px; vertical-align: top; border: 1px solid rgb(238, 238, 238);" >NodeManager</td><td style="box-sizing: border-box; padding: 8px; line-height: 20px; vertical-align: top; border: 1px solid rgb(238, 238, 238);" >NodeManager</td></tr></tbody></table><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >2 、 创建HDFS Federation 版本Hadoop程序目录</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >在bigdata01上创建目录/opt/modules/hadoopfederation /用来存放Hadoop Federation环境。</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 modules]$ mkdir /opt/modules/hadoopfederation
3、 新解压Hadoop 2.5.0
[hadoop@bigdata -senior01 ~]$ tar -zxf /opt/sofeware/hadoop-2.5 .0 .tar.gz -C /opt/modules/hadoopfederation/
4、 配置Hadoop JDK路径
修改hadoop-env.sh、mapred-env.sh、yarn-env.sh文件中的JDK路径。
export JAVA_HOME=”/opt/modules/jdk1.7.0_67”
5、 配置hdfs-site.xml
< configuration >
< property >
< !—配置三台NameNode -- >
< name > dfs.nameservices</ name >
< value > ns1,ns2,ns3</ value >
</ property >
< property >
< !—第一台NameNode的机器名和rpc端口,指定了NameNode和DataNode通讯用的端口号 -- >
< name > dfs.namenode.rpc-address.ns1</ name >
< value > bigdata-senior01.chybinmy.com:8020</ value >
</ property >
< property >
< !—第一台NameNode的机器名和rpc端口,备用端口号 -- >
< name > dfs.namenode.serviceerpc-address.ns1</ name >
< value > bigdata-senior01.chybinmy.com:8022</ value >
</ property >
< property >
< !—第一台NameNode的http页面地址和端口号 -- >
< name > dfs.namenode.http-address.ns1</ name >
< value > bigdata-senior01.chybinmy.com:50070</ value >
</ property >
< property >
< !—第一台NameNode的https页面地址和端口号 -- >
< name > dfs.namenode.https-address.ns1</ name >
< value > bigdata-senior01.chybinmy.com:50470</ value >
</ property >
<property >
<name > dfs.namenode.rpc-address.ns2</name >
<value > bigdata-senior02.chybinmy.com:8020</value >
</property >
<property >
<name > dfs.namenode.serviceerpc-address.ns2</name >
<value > bigdata-senior02.chybinmy.com:8022</value >
</property >
<property >
<name > dfs.namenode.http-address.ns2</name >
<value > bigdata-senior02.chybinmy.com:50070</value >
</property >
<property >
<name > dfs.namenode.https-address.ns2</name >
<value > bigdata-senior02.chybinmy.com:50470</value >
</property >
<property >
<name > dfs.namenode.rpc-address.ns3</name >
<value > bigdata-senior03.chybinmy.com:8020</value >
</property >
<property >
<name > dfs.namenode.serviceerpc-address.ns3</name >
<value > bigdata-senior03.chybinmy.com:8022</value >
</property >
<property >
<name > dfs.namenode.http-address.ns3</name >
<value > bigdata-senior03.chybinmy.com:50070</value >
</property >
<property >
<name > dfs.namenode.https-address.ns3</name >
<value > bigdata-senior03.chybinmy.com:50470</value >
</property >
</configuration >
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63
6、 配置core-site.xml
< configuration >
< property >
< name > hadoop.tmp.dir</ name >
< value > /opt/modules/hadoopha/hadoop-2.5.0/data/tmp</ value >
</ property >
</ configuration >
hadoop.tmp.dir设置hadoop临时目录地址,默认时,NameNode和DataNode的数据存在这个路径下。
7、 配置slaves文件
bigdata-senior01.chybinmy.com
bigdata-senior02.chybinmy.com
bigdata-senior03.chybinmy.com
8、 配置yarn-site.xml
< configuration >
< property >
< name > yarn.nodemanager.aux-services</ name >
< value > mapreduce_shuffle</ value >
</ property >
< property >
< name > yarn.resourcemanager.hostname</ name >
< value > bigdata-senior02.chybinmy.com</ value >
</ property >
< property >
< name > yarn.log-aggregation-enable</ name >
< value > true</ value >
</ property >
< property >
< name > yarn.log-aggregation.retain-seconds</ name >
< value > 106800</ value >
</ property >
</ configuration >
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 9、 分发到其他节点
分发之前先将share/doc目录删除,这个目录中是帮助文件,并且很大,可以删除。
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ scp -r /opt/modules/ /opt/modules/hadoopfederation bigdata-senior02.chybinmy.com:/opt/modules
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ scp -r /opt/modules/hadoopfederation bigdata-senior03.chybinmy.com:/opt/modules</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >10 、 格式化NameNode</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >在第一台上进行NameNode格式化。</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ bin/hdfs namenode -format -clusterId hadoop-federation-clusterId
这里一定要指定一个集群ID,使得多个NameNode的集群ID是一样的,因为这三个NameNode在同一个集群中,这里集群ID为hadoop-federation-clusterId。
在第二台NameNode上。
[hadoop@bigdata -senior02 hadoop-2.5 .0 ]$ bin/hdfs namenode -format -clusterId hadoop-federation-clusterId</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >在第二台NameNode上。</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior03 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ bin/hdfs namenode -format -clusterId hadoop-federation-clusterId
11、 启动NameNode
在第一台、第二台、第三台机器上启动NameNode:
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh start namenode
[hadoop@bigdata -senior02 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh start namenode
[hadoop@bigdata -senior03 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh start namenode</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >3 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >启动后,用jps命令查看是否已经启动成功。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >查看HDFS Web页面,此时三个NameNode都是standby状态。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/dd3582b0-5cbf-11e7-8185-21ba04c77532" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/e55b6310-5cbf-11e7-86d9-f17e4b747fa0" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" ><img src="http://images.gitbook.cn/ed9121f0-5cbf-11e7-8185-21ba04c77532" alt="enter image description here" title="" style="border: none; box-sizing: border-box; max-width: 100%;" ></p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >12 、 启动DataNode</p><pre class="prettyprint" name="code" style="white-space: nowrap; word-wrap: break-word; box-sizing: border-box; position: relative; overflow-y: hidden; overflow-x: auto; margin-top: 0px; margin-bottom: 1.1em; font-family: "Source Code Pro", monospace; padding: 5px 5px 5px 60px; font-size: 14px; line-height: 1.45; word-break: break-all; color: rgb(51, 51, 51); background-color: rgba(128, 128, 128, 0.05); border: 0px solid rgb(136, 136, 136); border-radius: 0px;" ><code class="language-java hljs has-numbering" style="display: block; padding: 0px; background: transparent; color: inherit; box-sizing: border-box; font-family: "Source Code Pro", monospace;font-size:undefined; white-space: pre; border-radius: 0px; word-wrap: normal;" >[hadoop<span class="hljs-annotation" style="color: rgb(155, 133, 157); box-sizing: border-box;" >@bigdata </span>-senior01 hadoop-<span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >2.5 </span><span class="hljs-number" style="color: rgb(0, 102, 102); box-sizing: border-box;" >.0 </span>]$ sbin/hadoop-daemon.sh start datanode
[hadoop@bigdata -senior02 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh start datanode
[hadoop@bigdata -senior03 hadoop-2.5 .0 ]$ sbin/hadoop-daemon.sh start datanode</code><ul class="pre-numbering" style="box-sizing: border-box; position: absolute; width: 50px; background-color: rgb(238, 238, 238); top: 0px; left: 0px; margin: 0px; padding: 6px 0px 40px; border-right: 1px solid rgb(221, 221, 221); list-style: none; text-align: right;" ><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >1 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >2 </li><li style="box-sizing: border-box; padding: 0px 5px; list-style: none; margin-left: 0px;" >3 </li></ul></pre><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >启动后,用jps命令确认DataNode进程已经启动成功。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >四十九、测试HDFS Federation</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >1 、 修改core-site.xml</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >在bigdata-senior01机器上,修改core-site.xml文件,指定连接的NameNode是第一台NameNode。</p><p style="margin-top: 0px; margin-bottom: 1.1em; padding-top: 0px; padding-bottom: 0px; box-sizing: border-box; color: rgb(63, 63, 63); font-family: "microsoft yahei"; font-size: 15px;" >[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ vim etc/hadoop/core-site.xml
< configuration >
< property >
< name > fs.defaultFS</ name >
< value > hdfs://bigdata-senior01.chybinmy.com:8020</ value >
</ property >
< property >
< name > hadoop.tmp.dir</ name >
< value > /opt/modules/hadoopfederation/hadoop-2.5.0/data/tmp</ value >
</ property >
</ configuration >
2、 在bigdate-senior01上传一个文件到HDFS
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ bin/hdfs dfs -mkdir /tmp
[hadoop@bigdata -senior01 hadoop-2.5 .0 ]$ bin/hdfs dfs -put ~/shuffle_daily.sh /tmp/shuffle_daily.sh
3、 查看HDFS文件
可以看到,刚才的文件只上传到了bigdate-senior01机器上的NameNode上了,并没有上传到其他的NameNode上去。
这样,在HDFS的客户端,可以指定要上传到哪个NameNode上,从而来达到了划分NameNode的目的。
后记这篇文章的操作步骤并不是工作中标准的操作流程,如果在成百上千的机器全部这样安装会被累死,希望读者可以通过文章中一步步地安装,从而初步了解到Hadoop的组成部分,协助过程等,这对于Hadoop的深入使用有很大的帮助。