Shell脚本完成hadoop的集群安装

虽然整体实现的自动安装，但还是有很多需要完善的地方，比如说：

1. 代码目前只能在root权限下运行，否则会出错，这方面需要加权限判断；

2.另外可以增加几个函数，减少代码冗余；

3.还有一些判断不够智能；

......

苦于能力和时间都有限，只能写到这里了。

installHadoop文件代码如下：

#!/bin/bash
#

root_password="123456"
jdk_tar=jdk-8u65-linux-i586.tar.gz
jdk_url=http://download.oracle.com/otn-pub/java/jdk/8u65-b17/jdk-8u65-linux-i586.tar.gz
jdk_version=jdk1.8.0_65
java_version=1.8.0_65
jdk_install_path=/usr/local/development
hadoop_url=http://101.44.1.4/files/2250000004303235/mirrors.hust.edu.cn/apache/hadoop/common/stable1/hadoop-1.2.1.tar.gz
hadoop_version=hadoop-1.2.1
hadoop_tar=hadoop-1.2.1.tar.gz
hadoop_install_path=hadoop
hadoop_tmp_path=/home/hadoop/hadoop_tmp
hadoop_name_path=/home/hadoop/hdfs/name
hadoop_data_path=/home/hadoop/hdfs/data
user_name=hadoop
user_passwd=hadoop

#su
#判断能否root
#if [ $? -ne 0 ] ;then
#	echo "No root access"
#	exit
#fi

shFilePath=$(pwd)

#check jdk installed or not
java -version &> /dev/null
if [ $? -ne 0 ] ;then
	echo {Jdk has been installed in this pc}
	java -version
else
	#检查~/../usr/local/development目录存在否，不存在就创建
	#先进入当前用户的家目录
	#cd ~
	#cd ../../usr/local/$jdk_install_path &> /dev/null
	#if [ $? -ne 0 ] ;then
        if [ ! -d $jdk_install_path ] ;then
		echo "{Create $jdk_install_path folder to install jdk}"
		mkdir $jdk_install_path
		cd $jdk_install_path
		echo "{Success to create $jdk_install_path folder}"
	else
		echo "{$jdk_install_path folder has already exists}"
		cd $jdk_install_path
	fi

	#检查jdk是否解压
	#ls | grep "$jdk_version" &> /dev/null
	if [ ! -d $jdk_version  ] ;then
		#检查jdk是否已有压缩包
		if [ ! -f $jdk_tar ] ;then
			echo "{Download $jdk_tar}"
			wget --no-check-certificate --no-cookies --header "Cookie: oraclelicense=accept-securebackup-cookie" $jdk_url
		fi
		echo "{Untar $jdk_tar}"
		tar -zxvf $jdk_tar

	else
		echo "{$jdk_version folder has already exists in $jdk_install_path/}"
	fi

	#set jdk environment
	echo {set java environment}
	cd ~
	cd ../../../../etc/profile.d/
	touch $jdk_install_path.sh
	#echo "#!bin/bash" > $jdk_install_path.sh
	echo "export JAVA_HOME=/usr/local/$jdk_install_path/$jdk_version" >> $jdk_install_path.sh
	echo "export JRE_HOME=\$JAVA_HOME/jre" >> $jdk_install_path.sh
	echo "export CLASSPATH=.:\$JAVA_HOME/lib:\$JRE_HOME/lib:\$CLASSPATH" >> $jdk_install_path.sh
	echo "PATH=\$JAVA_HOME/bin:\$JRE_HOME/bin:\$PATH"	>> $jdk_install_path.sh
      source $jdk_install_path.sh

	#check the java version
	java -version | grep "$java_version" &> /dev/null
	if [ $? -ne 0 ] ;then
		echo "{Success to install $jdk_version}"
	fi
fi

#no passwd when login via ss
echo "{Config ssh service and login via ssh without no passwd}"
sudo yum -y install ssh openssh-server
#update /etc/ssh/sshd_config
#RSAAuthentication
RSAAuthentication_lineNum=`awk '/RSAAuthentication yes/{print NR}' ~/../etc/ssh/sshd_config`
RSAAuthentication="RSAAuthentication yes"
sed -i "${RSAAuthentication_lineNum}s/^.*/${RSAAuthentication}/g" ~/../etc/ssh/sshd_config

#PubkeyAuthentication
PubkeyAuthentication_lineNum=`awk '/PubkeyAuthentication yes/{print NR}' ~/../etc/ssh/sshd_config`
PubkeyAuthentication="PubkeyAuthentication yes"
sed -i "${PubkeyAuthentication_lineNum}s/^.*/${PubkeyAuthentication}/g" ~/../etc/ssh/sshd_config

#AuthorizedKeysFile
AuthorizedKeysFile_lineNum=`awk '/AuthorizedKeysFile/{print NR}' ~/../etc/ssh/sshd_config`
AuthorizedKeysFile="AuthorizedKeysFile .ssh\/authorized_keys"
sed -i "${AuthorizedKeysFile_lineNum}s/^.*/${AuthorizedKeysFile}/g" ~/../etc/ssh/sshd_config

echo "{You change in sshd_config as follow}"
sed -n  "${RSAAuthentication_lineNum},${AuthorizedKeysFile_lineNum}p" ~/../etc/ssh/sshd_config

#restart sshd service
~/../sbin/service sshd restart
echo "{Finish to update sshd_config}"

#generate public key
if [ ! -d ~/.ssh ] ;then
	mkdir ~/.ssh
fi

cd ~/.ssh
echo y | ssh-keygen -t rsa -P '' -f id_rsa
if [ ! -f authorized_keys ] ;then
	touch authorized_keys
	cat id_rsa.pub > authorized_keys
else
	cat id_rsa.pub >> authorized_keys

fi

chmod 700 ~/.ssh
chmod 600 ~/.ssh/authorized_keys

#Download hadoop
cd ~
cd ../home/$hadoop_install_path &> /dev/null
if [ $? -ne 0 ] ;then
	echo "{Create /home/$hadoop_install_path folder to install jdk}"
	cd ../home
	mkdir $hadoop_install_path
	cd $hadoop_install_path
	echo "{Success to create $$hadoop_install_path folder}"
else
	echo "{/home/$hadoop_install_path folder has already exists}"
	cd ~
	cd ../home/$hadoop_install_path
fi

#check hadoop-2.7.0 folder is exists or not
if [ ! -d "$hadoop_version" ] ;then
	#check hadoop-2.7.0.tar.gz is exist or not
	if [ ! -f "$hadoop_tar" ] ;then
		echo "{Download $hadoop_tar}"
		wget --no-check-certificate --no-cookies --header "Cookie: oraclelicense=accept-securebackup-cookie" $hadoop_url
	fi
	echo "{Untar $hadoop_tar}"
	tar -zxvf $hadoop_tar
else
	echo "{$hadoop_version folder has already exists in /home/$hadoop_install_path/}"
fi

#enter into config folder
cd $hadoop_version
if [ ! -d "conf" ] ;then
	cd etc/hadoop/
else
	cd conf
fi

#update hadoop-env.sh
java_home_line_num=`awk '/export JAVA_HOME/{print NR}' hadoop-env.sh`
JAVAHOME="export JAVA_HOME=\/usr\/local\/"$jdk_install_path"\/"$jdk_version

#-i is directly modify the source file
sed -i "${java_home_line_num}s/^.*/${JAVAHOME}/g" hadoop-env.sh
cat hadoop-env.sh | grep "JAVA_HOME"
echo "{Finish to update hadoop-env.sh}"

hadoop_config_path=$(pwd)
#echo $cur_path
#echo $shFilePath
#unalias cp
#cp -rf core-site.xml $curPath/

cd $shFilePath

#update core_site.xml
cat core-site.xml > $hadoop_config_path/core-site.xml
if [ ! -d $hadoop_tmp_path ] ;then
	mkdir $hadoop_tmp_path
fi
rm -rf $hadoop_tmp_path/*

if [ ! -d $hadoop_name_path ] ;then
	mkdir $hadoop_name_path
fi
chmod g-w $hadoop_name_path
rm -rf $hadoop_name_path/*

if [ ! -d $hadoop_data_path ] ;then
	mkdir $hadoop_data_path
fi
chmod g-w $hadoop_data_path
rm -rf $hadoop_data_path/*

#update mapred-site.xml
cat mapred-site.xml > $hadoop_config_path/mapred-site.xml

#update hdfs-site.xml
cat hdfs-site.xml > $hadoop_config_path/hdfs-site.xml

cd $hadoop_config_path
echo "{Check core-site.xml}"
#cat core-site.xml
echo "{Check mapred-site.xml}"
#cat mapred-site.xml
echo "{Check hdfs-site.xml}"
#cat hdfs-site.xml
echo "{Finish config hadoop}"

#add hadoop account and has admin access
id $user_name
if [ $? -ne 0 ] ;then
	echo "{add $user_name}"
	sudo useradd -mr $user_name
fi
#set passwd for hadoop account
echo $user_passwd | sudo passwd  --stdin $user_name

echo "{Format hadoop}"
echo Y | ../bin/hadoop namenode -format
cd ../bin/
bash stop-all.sh
echo "{Start hadoop}"
bash start-all.sh

result=`jps | awk '{print $2}' | xargs`
expect_result="JobTracker NameNode DataNode TaskTracker Jps SecondaryNameNode"
if [ "$result" == "$expect_result" ] ;then
	echo "{Congratulations!!! Success to intall hadoop!}"
else
	echo "{Sorry, fail to install hadoop and try to restart hadroop!}"
	bash stop-all.sh
	echo "{Start hadoop}"
	bash start-all.sh
	result=`jps | awk '{print $2}' | xargs`
	if [ "$result" == "$expect_result" ] ;then
		echo "{Sorry, fail to find all java thread and please check!}"
	else
		echo "{Congratulations!!! find all java thread, success to install hadoop!}"
	fi
fi

echo {!!!finish!!!}

　　

此外为了实现自动化配置hadoop, 还需要把core-site.xml, hdfs-site.xml和mapred-site.xml文件放到与installHadoop文件同级目录下。

core-site.xml文件：

<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
<!-- Put site-specific property overrides in this file. -->
<configuration>
        <property>
            <name>fs.default.name</name>
            <value>hdfs://localhost:9000</value>
        </property>
    <property>
          <name>hadoop.tmp.dir</name>
          <value>/home/hadoop/hadoop_tmp</value>
         <description>A base for other temporary directories.</description>
    </property>
</configuration>

hdfs-site.xml:

<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
<!-- Put site-specific property overrides in this file. -->
<configuration>
<property>
    <name>dfs.name.dir</name>
    <value>/home/hadoop/hdfs/name</value>
</property>
<property>
    <name>dfs.data.dir</name>
    <value>/home/hadoop/hdfs/data</value>
</property>
<property>
    <name>dfs.replication</name>
    <value>1</value>
</property>
</configuration>

mapred-site.xml

<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
<!-- Put site-specific property overrides in this file. -->
<configuration>
    <property>
        <name>mapred.job.tracker</name>
        <value>localhost:9001</value>
    </property>
</configuration>

单机版的Hadoop安装可以参考：

http://www.linuxidc.com/Linux/2015-04/116447.htm

多台机器的Hadoop安装可以参考：

http://blog.csdn.net/ab198604/article/details/8250461