有Bug 欢迎反馈，我不烦：feiyuanxing@gmail.com
1 #!/bin/bash
2 #@author：feiyuanxing 【既然笨到家，就要努力到家】
3 #@date：2017-01-05
4 #@E-Mail：feiyuanxing@gmail.com
5 #@TARGET:一键安装hadoop 2.7.6 centos 64位
6 #@CopyRight:本脚本遵守 未来星开源协议（http://feiyuanxing.com/kaiyuanxieyi/kaiyuanxieyi.html）
7
8 #讲解：
9 #0.前提条件：
10 # 0.1 已经配置好JAVA环境变量
11 # 0.2 已经配置免密登录环境
12 #1.请使用root用户执行，软件默认安装在work用户，通常会在linux的/home 挂载一个大型磁盘
13 #2.软件安装在/home/work/local/hadoop/hadoop 有人问为什么这么深
14 # 2.1 work目录下local包含所有软件 hadoop安装大数据生态圈其他软件
15 # 2.2 hadoop的数据安装在 /home/work/data ，此目录由于数量较大，可以单独挂载一个磁盘
16 #
17 ####################################################################################
18 # 个性化设置
19 # Hadoop 生态目录
20 INSTALL_HOME=/home/work/local/hadoop
21
22 #data 目录
23 DATA_HOME=/home/work/data
24 # Hadoop目录
25 ROOT_HOME=${INSTALL_HOME}/hadoop_tmp
26
27 #name 通常为主机（本机），缺省配置为master
28 NAMENODE=master
29 #name 通常为第一台从机，缺省配置为slave01
30 SECONDARYNAMENODE=slave01
31 #hadoop 版本
32 HADOOP_VERSION=hadoop-2.7.6.tar.gz
33 ####################################################################################
34 # Hadoop 生态数据路径
35 hadoop_logs=${DATA_HOME}/hadoop_logs
36 hadoop_tmp=${DATA_HOME}/hadoop_tmp
37 hadoop_data=${DATA_HOME}/hadoop_data
38 hadoop_name=${DATA_HOME}/hadoop_name
39 #当前脚本的位置
40 SCRIPT_HOME=$(cd `dirname $0`; pwd)
41
42 mkdir -p ${INSTALL_HOME} && cd ${INSTALL_HOME}
43 mkdir -p ${ROOT_HOME} && mkdir -p ${hadoop_logs} && mkdir -p ${hadoop_tmp}&& mkdir -p ${hadoop_data}&& mkdir -p ${hadoop_name}
44
45 #chown -R work:work /home/work/data/hadoop
46
47 #集群机器信息
48 INSTALL_IP_ARRAY=()
49 INSTALL_HOSTNAME_ARRAY=()
50
51
52 # 需要判断用户是否存在
53 function add_work_user(){
54 adduer work -d /home/work
55 passwd work
56
57 }
58
59 #需要判断jdk是否存在
60 function install_jdk(){
61 # 有jdk
62 echo $JAVA_HOME
63
64 }
65
66
67 #下载hadoop
68 function download_hodoop(){
69 echo "(2-1/6)正在获取安装包：${INSTALL_HOME}/hadoop-2.7.6.tar.gz"
70 if [ ! -f ${INSTALL_HOME}/hadoop-2.7.6.tar.gz ] ;then
71 echo "开始下载"
72 wget https://mirrors.tuna.tsinghua.edu.cn/apache/hadoop/common/hadoop-2.7.6/${HADOOP_VERSION}
73 fi
74 echo "(2-2/6)获取安装包完成：${INSTALL_HOME}/hadoop-2.7.6.tar.gz"
75
76 }
77
78 function configuration_ssh(){
79 #设置本机免密码登录
80 (echo -e "\n"
81 sleep 1
82 echo -e "\n"
83 sleep 1
84 echo -e "\n")|ssh-keygen -t rsa
85 cat ~/.ssh/id_rsa.pub >> ~/.ssh/authorized_keys
86 chmod 700 ~/.ssh
87 chmod 600 ~/.ssh/authorized_keys
88 }
89
90 function readFileToArra(){
91 echo "(1/6)正在读取配置文件：${SCRIPT_HOME}/hostip_hadoop.txt"
92 if [ ! -f "${SCRIPT_HOME}"/hostip_hadoop.txt ];then
93 echo "请在本目录下的【hostip_hadoop.txt】文件中添加Hadoop集群机器信息"
94 echo "#USENAME@IP:PORT,PASSWD（第一个必须配置为主，通常为本机）
95 root@191.168.1.1:22,123456
96 root@191.168.1.2:22,123456" > ${SCRIPT_HOME}/hostip_hadoop.txt
97 exit;
98 else
99 INSTALL_IP_ARRAY=($(cat "${SCRIPT_HOME}"/hostip_hadoop.txt| grep -v "^#"))
100 INSTALL_HOSTNAME_ARRAY=($(cat "${SCRIPT_HOME}"/hostip_hadoop.txt| grep -v "^#"|awk -F '@' '{print $2}'|awk -F ':' '{print $1}'))
101 fi
102
103 #for(( i=0 ; i<${#INSTALL_IP_ARRAY[@]} ; i++)) do
104 # echo "打印数组：${INSTALL_IP_ARRAY[i]}"
105 #done;
106 }
107
108 #readFileToArra
109
110 function configure_hadoop(){
111
112 echo "(3-1/6)正在解压安装包："
113 #最终生成的文件名为install_hadoop.bin而我们的all.tar.gz被>>到该文件后面
114 #tail -c $size install_hadoop.bin >all.tar.gz
115 rm -rf hadoop hadoop_tmp
116 tar -zxf ${HADOOP_VERSION}
117 #pwd
118 mv hadoop-2.7.6 hadoop_tmp && cd hadoop_tmp
119
120 echo "(3-2/6)正在配置Hadoop环境变量："
121 #设置环境变量
122 hadoop_home_before=`cat /etc/profile| grep "HADOOP_HOME"`
123
124 if [ -z "${hadoop_home_before}" ] ;then
125 echo "hadoop 环境变量已存在..."
126 sed -i '/HADOOP_HOME/d' /etc/profile
127 echo "export HADOOP_HOME=${INSTALL_HOME}" >> /etc/profile
128 echo 'export PATH=$PATH:${HADOOP_HOME}/bin:${HADOOP_HOME}/sbin' >> /etc/profile
129 #使得环境变量生效
130 source /etc/profile
131 sleep 1;
132 source /etc/profile
133
134 fi
135
136 echo "(4/6)正在配置Hadoop基本配置："
137 #通过强大的sed命令来修改hadoop的配置文件
138 jdk_home=`echo $JAVA_HOME`
139 if [ -z ${jdk_home} ] ;then
140 #安装JDK
141 echo "程序已退出，请先配置JDK"
142 exit
143 fi
144
145 #echo "开始配置hadoop"
146 # 1.配置hadoop-env.sh
147 sed -i "s!\${JAVA_HOME}!$(echo ${jdk_home})!g" ${ROOT_HOME}/etc/hadoop/hadoop-env.sh
148 # 2.配置core-site.xml
149 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>fs.defaultFS\<\/name\>\n\\t\\t\<value\>hdfs://${NAMENODE}:9000\<\/value\>\n\\t\<\/property\>\n\\t\<property\>\n\\t\\t\<name\>hadoop.tmp.dir\<\/name\>\n\\t\\t\<value\>file:$(echo ${hadoop_tmp})\<\/value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/core-site.xml
150 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>io.file.buffer.size\</name\>\n\\t\\t\<value\>131072\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/core-site.xml
151 # 3.配置hdfs-site.xml
152 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>dfs.webhdfs.enabled\</name\>\n\\t\\t\<value\>true\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/hdfs-site.xml
153 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>dfs.replication\</name\>\n\\t\\t\<value\>2\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/hdfs-site.xml
154 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>dfs.datanode.data.dir\</name\>\n\\t\\t\<value\>file:${hadoop_name}\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/hdfs-site.xml
155 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>dfs.namenode.name.dir\</name\>\n\\t\\t\<value\>file:${hadoop_data}\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/hdfs-site.xml
156 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>dfs.namenode.secondary.http-address\</name\>\n\\t\\t\<value\>master:9001\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/hdfs-site.xml
157 # 4.配置mapred-site.xml
158 cp ${ROOT_HOME}/etc/hadoop/mapred-site.xml.template ${ROOT_HOME}/etc/hadoop/mapred-site.xml
159 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>mapreduce.jobhistory.webapp.address\</name\>\n\\t\\t\<value\>master:19888\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/mapred-site.xml
160 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>mapreduce.jobhistory.address\</name\>\n\\t\\t\<value\>master:10020\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/mapred-site.xml
161 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>mapreduce.framework.name\</name\>\n\\t\\t\<value\>yarn\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/mapred-site.xml
162 # 5.配置yarn-site.xml
163 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>yarn.resourcemanager.webapp.address\</name\>\n\\t\\t\<value\>master:8088\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/yarn-site.xml
164 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>yarn.resourcemanager.admin.address\</name\>\n\\t\\t\<value\>master:8033\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/yarn-site.xml
165 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>yarn.resourcemanager.resource-tracker.address\</name\>\n\\t\\t\<value\>master:8031\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/yarn-site.xml
166 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>yarn.resourcemanager.scheduler.address\</name\>\n\\t\\t\<value\>master:8030\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/yarn-site.xml
167 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>yarn.resourcemanager.address\</name\>\n\\t\\t\<value\>master:8032\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/yarn-site.xml
168 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>yarn.nodemanager.aux-services.mapreduce.shuffle.class\</name\>\n\\t\\t\<value\>org.apache.hadoop.mapred.ShuffleHandler\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/yarn-site.xml
169 sed -i "/<configuration>/a\\\t\<property\>\n\\t\\t\<name\>yarn.nodemanager.aux-services\</name\>\n\\t\\t\<value\>mapreduce_shuffle\</value\>\n\\t\</property\>" ${ROOT_HOME}/etc/hadoop/yarn-site.xml
170 # 6.配置slaves文件
171 echo "" > ${ROOT_HOME}/etc/hadoop/slaves
172 for(( i=0 ; i<${#INSTALL_HOSTNAME_ARRAY[@]} ; i++)) do
173 echo "${INSTALL_HOSTNAME_ARRAY[i]}" >> ${ROOT_HOME}/etc/hadoop/slaves
174 done;
175
176 echo "(5/6)正在发送hadoop到其他机器"
177 for line in ${INSTALL_IP_ARRAY[@]}
178 do
179 user=`echo ${line}| awk -F "@" '{print $1}'`
180 ip=`echo ${line}| cut -d "@" -f 2| cut -d ":" -f 1`
181 port=`echo ${line}| cut -d "@" -f 2| cut -d ":" -f 2|cut -d "," -f 1`
182 echo ${user} - ${ip} - ${port} - ${passwd}
183 ssh -p ${port} -o StrictHostKeyChecking=no "${user}@${ip}" "rm -rf ${DATA_HOME}/hadoop_*;rm -rf ${INSTALL_HOME}/hadoop;mkdir -p ${INSTALL_HOME}/hadoop;mkdir -p ${hadoop_logs} && mkdir -p ${hadoop_tmp}&& mkdir -p ${hadoop_data}&& mkdir -p ${hadoop_name}"
184 #echo "scp -P ${port} -o StrictHostKeyChecking=no -rpq ${ROOT_HOME} \"${user}\"@\"${ip}\":${INSTALL_HOME}"
185 scp -P ${port} -o StrictHostKeyChecking=no -rpq ${ROOT_HOME}/* "${user}"@"${ip}":${INSTALL_HOME}/hadoop
186 ssh -p ${port} -o StrictHostKeyChecking=no "${user}@${ip}" "chown -R work:work ${INSTALL_HOME}"
187 done
188
189 echo "(6/6)Hadoop配置成功"
190
191 }
192
193 #配置Hadoop启动服务
194 function start_hadoop(){
195 #格式化hadoop
196 hdfs namenode -format
197 echo "格式化完成..开始运行"
198 #先启动HDFS
199 start-dfs.sh
200 #再启动YARN
201 start-yarn.sh
202
203 #检查是否启动成功
204 jps
205
206 }
207 #动态读取配置文件
208 readFileToArra
209 #add_work_user
210 download_hodoop;
211 #configuration_ssh
212 configure_hadoop
213 #start_hadoop
