hadoop2.2.0源代码编译
hadoop2.2.0源代码编译一、环境说明
虚拟软件:VMware Workstation 10
虚拟机配置:
RHEL Server> cpu:4核心,内存:4G,硬盘:50G
二、前提条件:
1:将rhel6.5的iso文件作为yum源
2:hadoop-2.2.0-src.tar.gz
3:安装JDK 1.6.0_43
4:安装并配置apache-maven 3.0.5(apache-maven-3.0.5-bin.tar.gz)
源码中BUILDING.txt中要求使用3.0，从hadoop2.0版本以后使用maven编译，之前用Ant)
解压并配置环境变量
mvn -version
5:安装并配置apache-ant-1.9.3-bin.zip（下载二进制版本的,这个需要编译安装findbugs)
解压并配置环境变量
ant -version
6:下载并安装cmake cmake-2.8.12.1,安装命令如下:
tar -zxvf cmake-2.8.12.1.tar.gz
cd cmake-2.8.12.1
./bootstrap
make
make install
检查安装是否正确
cmake --version(如果能正确显示版本号,则说明安装正确)
7:下载并安装配置findbugs-2.0.2-source.zip
http://sourceforge.jp/projects/sfnet_findbugs/releases/
使用ant编译安装。如果不编译安装则编译的时候会报
hadoop-common-project/hadoop-common/${env.FINDBUGS_HOME}/src/xsl/default.xsl doesn’t exist. ->
进入到解压后的目录，直接运行ant命令
如果不安装,则在编译时会报如下错误:
Failed to execute goal org.apache.maven.plugins:maven-antrun-plugin:1.6:run (site) on project hadoop-common: An Ant BuildException has occured
8:安装zlib-devel
默认情况下,系统没有安装zlib-devel
yum install zlib-devel
如果不安装,则在编译时会报如下错误:
Failed to execute goal org.apache.maven.plugins:maven-antrun-plugin:1.6:run (make) on project hadoop-common
9: protobuf-2.5.0
yum install gcc-c++ (如果不安装,则cmake configure失败)
./configure
make
make check
make install
检查安装是否正确
protoc --version((如果能正确显示版本号,则说明安装正确)
三、hadoop2.2源码编译
1：进入hadoop2.2.0解压后的源码目录
2：执行mvn命令编译，此过程需要连接网络，编译的速度取决于你的网速
mvn clean package -Pdist,native -DskipTests -Dtar
2.1.Create binary distribution without native code and without documentation:
$ mvn package -Pdist -DskipTests -Dtar
2.2.Create binary distribution with native code and with documentation:
$ mvn package -Pdist,native,docs -DskipTests -Dtar
2.3.Create source distribution:
$ mvn package -Psrc -DskipTests
2.4.Create source and binary distributions with native code and documentation:
$ mvn package -Pdist,native,docs,src -DskipTests -Dtar
2.5.Create a local staging version of the website (in /tmp/hadoop-site)
$ mvn clean site; mvn site:stage -DstagingDirectory=/tmp/hadoop-site
3：编译后的项目发布版本在hadoop-2.2.0-src/hadoop-dist/target/目录下
hadoop-2.2.0-src/hadoop-dist/target/hadoop-2.2.0/
四、安装hadoop 单点伪分布式模式
1：配置ssh互信
ssh-keygen -t dsa （或行ssh-keygen -t rsa -P "" 加上-P ""参数只需要一次回车就可以执行完毕。不加需要三次回车）
cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys
如果仍然需要输入密码，则执行如下命令
chmod 600 ~/.ssh/authorized_keys
2：将编译后的hadoop-2.2.0-src/hadoop-dist/target/hadoop-2.2.0复制到/data/hadoop/目录下
3：建议软链接 ln -s hadoop-2.2.0 hadoop2
4: 在用户的.bash_profile增加如下变量
export HADOOP_HOME=/data/hadoop/hadoop2
export PATH=$PATH:$HADOOP_HOME/bin
export PATH=$PATH:$HADOOP_HOME/sbin
export HADOOP_MAPRED_HOME=${HADOOP_HOME}
export HADOOP_COMMON_HOME=${HADOOP_HOME}
export HADOOP_HDFS_HOME=${HADOOP_HOME}
export YARN_HOME=${HADOOP_HOME}
5:建议data.dir及namenode.dir目录
mkdir hdfs
mkdir namenode
chmod -R 755 hdfs
6:修改hadoop-env.sh
export JAVA_HOME=/usr/java/jdk1.6.0_43
export HADOOP_HOME=/data/hadoop/hadoop2
7:修改core-site.xml,hdfs-site.xml,mapred-site.xml,yarn-site.xml，内容见附录2
8：格式化HDFS文件系统
执行命令 hadoop namenode -format
9：启动hdfs,yarn
start-dfs.sh
start-yarn.sh
10:验证启动是否成功
jps
如果显示如下6个进程，则启动成功
53244 ResourceManager
53083 SecondaryNameNode
52928 DataNode
53640 Jps
52810 NameNode
53348 NodeManager
五、运行自带的wordcount例子
hadoop fs -mkdir /tmp
hadoop fs -mkdir /tmp/input
hadoop fs -put /usr/hadoop/test.txt /tmp/input
cd /data/hadoop/hadoop2/share/hadoop/mapreduce
hadoop jar hadoop-mapreduce-examples-2.2.0.jar wordcount /tmp/input /tmp/output
如果能正确运行，则hadoop安装配置正确
六、附录1：
设置的环境变量（/etc/profile,编辑后运行source /etc/profile/ 使配置生效）
#java set
export JAVA_HOME=/usr/java/jdk1.6.0_43
export JRE_HOME=/usr/java/jdk1.6.0_43/jre
export> export PATH=$PATH:$JAVA_HOME/bin
#maven set
export M2_HOME=/home/soft/maven
export PATH=$PATH:$M2_HOME/bin
#ant
export ANT_HOME=/home/soft/apache-ant-1.9.3
export PATH=$PATH:$ANT_HOME/bin
#findbugs
export FINDBUGS_HOME=/home/soft/findbugs-2.0.2
export PATH=$PATH:$FINDBUGS_HOME/bin
附录2：core-site.xml,hdfs-site.xml,mapred-site.xml,yarn-site.xml的文件内容
core-site.xml内容：
fs.defaultFS
hdfs://vdata.kt:8020
hdfs-site.xml内容：
dfs.name.dir
/data/hadoop/namenode
dfs.data.dir
/data/hadoop/hdfs
dfs.replication
1
dfs.permissions
false
mapred-site.xml内容：
mapreduce.framework.name
yarn
yarn-site.xml内容：
yarn.nodemanager.aux-services
mapreduce_shuffle
页:
[1]