一、 基本环境搭建
1. 设置主机名、静态IP/DNS、主机映射、windows主机映射(方便ssh访问与IP修改)等
设置主机名: vi /etc/sysconfig/network # 重启系统生效(临时修改: hastname xxx;另起一个终端将会看到效果,需要注意的是: 若即将搭建Hadoop,这里起的hostname禁止使用"_")
设置静态IP/DNS: vi /etc/sysconfig/network-scripts/ifcfg-eth0(示例：修改BOOTPROTO=static;IPADDR=192.168.0.111;GATEWAY=192.168.0.1;DNS1=192.168.0.1,重启网络服务: service network restart)
设置主机映射: vi /etc/hosts (格式：IP 主机名(hostname))
设置window主机映射: 修改host文件,添加 [IP 主机名]
关闭防火墙:chkconfig iptables off/service iptables restart(临时修改: service iptables stop/start 立即生效)
关闭selinx:vi /etc/sysconfig/selinux # 需要重启系统生效(linux的一个加强安全子系统,加强对文件的访问控制,临时关闭(放开):setenforce 0;临时开启:setenforce 1)
查看linux中是否有自带的open jdk,有则卸载,以免后期和后面安装jdk冲突而不生效(查看是否存在: java -version,如果已存在则查看java版本: rpm -qa | grep "java",卸载 rpm -e "查出来的java版本" 或 yum -y remove "查出来的java版本")
准备压缩包:
hadoop-2.5.0.tar.gz
hadoop-2.5.0-src.tar.gz(可选包,编译源码包时使用)
native-2.5.0.tar.gz(可选包,已编译好的hadoop库,可直接替换使用)
protobuf-2.5.0.tar.gz(可选包,编译源码是必备组件)
jdk-7u67-linux-x64.tar.gz(hadoop2.x要求jdk版本1.7+)
apache-maven-3.0.5-bin.tar.gz(Maven包)
repository.tar.gz(可选包,Maven仓库,在编译Hadoop源码,会用到,若不用，则在编译时会花费更长时间去下载)
eclipse-jee-kepler-SR1-linux-gtk-x86_64.tar.gz(linux下使用,编写mr程序本地测试使用)
2. 添加好用户,建立文件夹,并将准备文件上传至files
[root@centos66-bigdata-hadoop ~]# su - liuwl
[liuwl@centos66-bigdata-hadoop ~]$ cd opt/
[liuwl@centos66-bigdata-hadoop opt]$ ls
data files localsrc modules software workspace
---------------------------------------------------------------
上传搭建Hadoop2.x的所有tar压缩包,压缩包自备,使用上传工具
上传工具很多:filezilla,FlashFXP,Xftp,vmware-tools,notepad++...
可能会有文件夹权限问题,需要检查一下
3. 创建用户分配权限liuwl,并使用visudo给liuwl
[root@centos66-bigdata-hadoop ~]# visudo
...
liuwl ALL=(root) NOPASSWD:ALL
[root@centos66-bigdata-hadoop ~]# su - liuwl
[liuwl@centos66-bigdata-hadoop ~]$ sudo -l
...
User liuwl may run the following commands on this host:
(root) NOPASSWD: ALL
4. 建立文件目录
[root@centos66-bigdata-hadoop ~]# su - liuwl
[liuwl@centos66-bigdata-hadoop ~]$ cd opt/
[liuwl@centos66-bigdata-hadoop opt]$ ls
data files localsrc modules software workspace # 文件夹随意,自己知道是装载什么的就好
5. 安装 jdk-7u67-linux-x64(注意jdk版本号和是合适的系统位数,我这里是CentOS_66_64)
[liuwl@centos66-bigdata-hadoop ~]$ vi /etc/profile
...
#JAVA_HOME
export JAVA_HOME=/opt/modules/jdk1.7.0_67
export PATH=$PATH:$JAVA_HOME/bin
[liuwl@centos66-bigdata-hadoop ~] source /etc/profile
[liuwl@centos66-bigdata-hadoop ~]$ echo $JAVA_HOME
/opt/modules/jdk1.7.0_67
[liuwl@centos66-bigdata-hadoop ~]$ 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)
6. 解压hadoop-2.5.0.tar.gz并删除doc文档(doc文件太大,且不常使用可拷出来日常查看)
# 有兴趣的朋友可以使用lynx在终端查看doc文档,当然需要使用root用户安装lynx：yum -y instatll lynx
# 然后lynx xxx.html 退出：q-->y
[liuwl@centos66-bigdata-hadoop ~]$ cd /home/liuwl/opt/files
[liuwl@centos66-bigdata-hadoop files]$ tar -zxf hadoop-2.5.0.tar.gz -C ../modules/
[liuwl@centos66-bigdata-hadoop files]$ sudo rm -rf ../modules/hadoop-2.5.0/share/doc/
二、 Hadoop伪分布模式搭建(正题)
★ 配置文件目录:/home/liuwl/opt/modules/hadoop-2.5.0/etc/hadoop
PS:使用notepad++(NppFTP,若没有自行下载该组件)
1. 为xxx.env.sh配置jdk,即JAVA_HOME
hadoop-env.sh
export JAVA_HOME=/opt/modules/jdk1.7.0_67
mapred-env.sh
export JAVA_HOME=/opt/modules/jdk1.7.0_67
yarn-env.sh
export JAVA_HOME=/opt/modules/jdk1.7.0_67
2. 配置Hadoop自定义文件
1> hdfs >>
▶ namenode >>
core-site.xml >>
<!--指定namenode主机地址-->
<property>
<name>fs.defaultFS</name>
<value>hdfs://centos66-bigdata-hadoop.com:8020</value>
</property> <!--指定hdfs格式化临时目录-->
<property>
<name>hadoop.tmp.dir</name>
<value>/home/liuwl/opt/modules/hadoop-2.5.0/data/tmp</value>
</property> <!--修改外部web访问的账户，更改dr.who为liuwl(自定义)-->
<property>
<name>hadoop.http.staticuser.user</name>
<value>liuwl</value>
</property>
▶ datanode >>
slaves >>
centos66-bigdata-hadoop.com
hdfs-site.xml >>
<!--设置系统快副本个数-->
<property>
<name>dfs.replication</name>
<value>1</value>
</property> <!--访问jar运行后的临时目录去除权限限制-->
<property>
<name>dfs.permissions.enabled</name>
<value>false</value>
</property>
2> 格式化hdfs >>
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ bin/hdfs namenode -format
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ ls data/tmp/
dfs
3> 配置Yarn环境(包括SecondaryNameNode,JobHistoryServer) >>
yarn-site.xml >>
<!--告知系统resourcemanager所在机器-->
<property>
<name>yarn.resourcemanager.hostname</name>
<value>centos66-bigdata-hadoop.com</value>
</property> <!--告知系统在nodemanager上运行MR程序-->
<property>
<name>yarn.nodemanager.aux-services</name>
<value>mapreduce_shuffle</value>
</property> <!--启用日志聚集功能-->
<property>
<name>yarn.log-aggregation-enable</name>
<value>true</value>
</property> <!--配置日志保存期限,单位为秒-->
<property>
<name>yarn.log-aggregation.retain-seconds</name>
<value>108600</value>
</property>
4> 配置mapreduce环境
mapred.site.xml >>
<!--指定MapReduce运行在YARN上-->
<property>
<name>mapreduce.framework.name</name>
<value>yarn</value>
</property> <!--配置historyserver指定机器-->
<property>
<name>mapreduce.jobhistory.address</name>
<value>centos66-bigdata-hadoop.com:10020</value>
</property> <!--配置web访问historyserver-->
<property>
<name>mapreduce.jobhistory.webapp.address</name>
<value>centos66-bigdata-hadoop.com:19888</value>
</property>
5> 分别启动
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ sbin/hadoop-daemon.sh start namenode
starting namenode, logging to /home/liuwl/opt/modules/hadoop-2.5.0/logs/hadoop-liuwl-namenode-centos66-bigdata-hadoop.com.out
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ sbin/hadoop-daemon.sh start datanode
starting datanode, logging to /home/liuwl/opt/modules/hadoop-2.5.0/logs/hadoop-liuwl-datanode-centos66-bigdata-hadoop.com.out
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ sbin/yarn-daemon.sh start resourcemanager
starting resourcemanager, logging to /home/liuwl/opt/modules/hadoop-2.5.0/logs/yarn-liuwl-resourcemanager-centos66-bigdata-hadoop.com.out
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ sbin/yarn-daemon.sh start nodemanager
starting nodemanager, logging to /home/liuwl/opt/modules/hadoop-2.5.0/logs/yarn-liuwl-nodemanager-centos66-bigdata-hadoop.com.out
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ sbin/mr-jobhistory-daemon.sh start historyserver
starting historyserver, logging to /home/liuwl/opt/modules/hadoop-2.5.0/logs/mapred-liuwl-historyserver-centos66-bigdata-hadoop.com.out
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ sbin/hadoop-daemon.sh start secondarynamenode
starting secondarynamenode, logging to /home/liuwl/opt/modules/hadoop-2.5.0/logs/hadoop-liuwl-secondarynamenode-centos66-bigdata-hadoop.com.out
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ jps
10772 NameNode
11179 NodeManager
10853 DataNode
10938 ResourceManager
11382 SecondaryNameNode
11302 JobHistoryServer
11420 Jps
3. 测试hdfs文件系统
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ bin/hdfs dfs -mkdir -p /user/liuwl/tmp
16/09/14 07:51:14 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ vi ../../data/wordcount.input
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ bin/hdfs dfs -mkdir -p /user/liuwl/tmp/input
16/09/14 07:54:17 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ bin/hdfs dfs -put ../../data/wordcount.input /user/liuwl/tmp/input
16/09/14 07:54:45 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ bin/hdfs dfs -cat /user/liuwl/tmp/input/wordcount.input
16/09/14 07:55:27 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
hadoop mapreduce
yarn historyserver hadoop
mapreduce yarn
namenode datanode
datanode
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ bin/hdfs dfs -get /user/liuwl/tmp/input/wordcount.input /opt/modules/wc.input
16/09/14 07:56:24 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
get: /opt/modules/wc.input._COPYING_ (Permission denied)
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ bin/hdfs dfs -get /user/liuwl/tmp/input/wordcount.input ~/opt/data/wc.input
16/09/14 07:57:00 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ cat ../../data/wc.input
hadoop mapreduce
yarn historyserver hadoop
mapreduce yarn
namenode datanode
datanode
4. 使用mapreduce运行jar文件
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ bin/yarn jar share/hadoop/mapreduce/hadoop-mapreduce-examples-2.5.0.jar wordcount /user/liuwl/tmp/input /user/liuwl/tmp/output
16/09/14 07:59:53 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
16/09/14 07:59:55 INFO client.RMProxy: Connecting to ResourceManager at centos66-bigdata-hadoop.com/192.168.0.110:8032
16/09/14 07:59:57 INFO input.FileInputFormat: Total input paths to process : 1
16/09/14 07:59:57 INFO mapreduce.JobSubmitter: number of splits:1
16/09/14 07:59:58 INFO mapreduce.JobSubmitter: Submitting tokens for job: job_1473864360962_0001
16/09/14 07:59:59 INFO impl.YarnClientImpl: Submitted application application_1473864360962_0001
16/09/14 08:00:00 INFO mapreduce.Job: The url to track the job: http://centos66-bigdata-hadoop.com:8088/proxy/application_1473864360962_0001/
16/09/14 08:00:00 INFO mapreduce.Job: Running job: job_1473864360962_0001
16/09/14 08:00:30 INFO mapreduce.Job: Job job_1473864360962_0001 running in uber mode : false
16/09/14 08:00:30 INFO mapreduce.Job: map 0% reduce 0%
16/09/14 08:01:19 INFO mapreduce.Job: map 100% reduce 0%
16/09/14 08:01:47 INFO mapreduce.Job: map 100% reduce 100%
16/09/14 08:01:49 INFO mapreduce.Job: Job job_1473864360962_0001 completed successfully
16/09/14 08:01:54 INFO mapreduce.Job: Counters: 49
File System Counters
FILE: Number of bytes read=96
FILE: Number of bytes written=194473
FILE: Number of read operations=0
FILE: Number of large read operations=0
FILE: Number of write operations=0
HDFS: Number of bytes read=226
HDFS: Number of bytes written=66
HDFS: Number of read operations=6
HDFS: Number of large read operations=0
HDFS: Number of write operations=2
Job Counters
Launched map tasks=1
Launched reduce tasks=1
Data-local map tasks=1
Total time spent by all maps in occupied slots (ms)=48483
Total time spent by all reduces in occupied slots (ms)=21661
Total time spent by all map tasks (ms)=48483
Total time spent by all reduce tasks (ms)=21661
Total vcore-seconds taken by all map tasks=48483
Total vcore-seconds taken by all reduce tasks=21661
Total megabyte-seconds taken by all map tasks=49646592
Total megabyte-seconds taken by all reduce tasks=22180864
Map-Reduce Framework
Map input records=5
Map output records=10
Map output bytes=125
Map output materialized bytes=96
Input split bytes=141
Combine input records=10
Combine output records=6
Reduce input groups=6
Reduce shuffle bytes=96
Reduce input records=6
Reduce output records=6
Spilled Records=12
Shuffled Maps =1
Failed Shuffles=0
Merged Map outputs=1
GC time elapsed (ms)=293
CPU time spent (ms)=2970
Physical memory (bytes) snapshot=313458688
Virtual memory (bytes) snapshot=1680084992
Total committed heap usage (bytes)=136450048
Shuffle Errors
BAD_ID=0
CONNECTION=0
IO_ERROR=0
WRONG_LENGTH=0
WRONG_MAP=0
WRONG_REDUCE=0
File Input Format Counters
Bytes Read=85
File Output Format Counters
Bytes Written=66
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ bin/hdfs dfs -ls /user/liuwl/tmp/output
16/09/14 08:02:23 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
Found 2 items
-rw-r--r-- 1 liuwl supergroup 0 2016-09-14 08:01 /user/liuwl/tmp/output/_SUCCESS
-rw-r--r-- 1 liuwl supergroup 66 2016-09-14 08:01 /user/liuwl/tmp/output/part-r-00000
[liuwl@centos66-bigdata-hadoop hadoop-2.5.0]$ bin/hdfs dfs -text /user/liuwl/tmp/output/part*
16/09/14 08:02:44 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
datanode 2
hadoop 2
historyserver 1
mapreduce 2
namenode 1
yarn 2
5. 简述hadoop四大组件原理
1> Hadoop Common:hadoop的公共类，方法，功能
2> Hadoop Distributed File System(hafs)
hadoop 分布式 文件系统
架构:主从架构(分工明确,namenode存储从节点信息,datanode存储具体数据)
可靠性:
系统块副本机制(自定义副本个数,坏块就近自动填补,定期校验副本块)
文件系统使用SecondaryNameNode定期合并edit与影像文件
可扩展性:
在集群全有机器基础上可任意添加多台机器
运行原理:
客户端写入文件,告知namenode,namenode存储着datanode以及以前文件的所有信息,分配系统块给予客户端写入
客户端读文件，namenode根据文件信息快速找到文件,采用就近原则,返回给用户
3> Hadoop Yarn:hadoop统一资源管理与任务调度框架
架构:主从架构(ResourceManager与NodeManager)
个人认为,yarn类似javaee中spring框架,作为了一个容器使用
yarn工作流程:客户端提交一个job,ResourceManager中ApplicationManager为job通过NodeManager建立ApplicationMaster用于管理job和反馈信息，ApplicationMaster告知ApplicationManager,所需要的所有正常运行job的资源,包括cpu,内存等,ApplicationManager返回给ApplicationMaster一个container(容器),让job在该容器中运行,其他job无法争夺其中的的资源,起到很好的隔离作用,job运行完毕会将运行信息发回给ApplicationMaster,ApplicationMaster通知ApplicationManager任务运行的情况,并记录job运行历史文件,收回资源等
4> Hadoop MapReduce:MapReduce是一个任务运行工具,每一个map便会开启一个java虚拟机,在MapReduceOnYarn时每个任务通过RPC协议向ApplicationManager报告自己的状态
