centos7 安装 hadoop-cdh-2.5 on mesos

1.环境说明

master 192.168.0.223 mesos-masterslave 192.168.0.225 mesos-salve2.环境准备
关闭防火墙
关闭selinux
两台机器修改主机名master/slave
设置hosts，能互相解析

3.master和slave配置ssh互信

这里配置hadoop用户双机互信，因为hadoop用hadoop用户启动

master
yum  -y install  sshpass
ssh-keygen 一路回车
ssh-copy-id -i ~/.ssh/id_rsa.pub hadoop@192.168.0.220
slave
yum  -y install  sshpass
ssh-keygen 一路回车
ssh-copy-id -i ~/.ssh/id_rsa.pub hadoop@192.168.0.201
测试ssh对方主机，不提示输入密码则OK

4.安装JDK

tar zxvf jdk-8u65-linux-x64.tar.gz
mv jdk1.8.0_65  /usr/jdk

4.1设置环境变量

export JAVA_HOME=/usr/jdk
export JRE_HOME=/usr/jdk/jre
export CLASSPATH=.:$CLASSPATH:$JAVA_HOME/lib:$JRE_HOME/lib
export PATH=$PATH:$JAVA_HOME/bin:$JRE_HOME/bin
执行 source  /etc/profile

4.2测试JDK

java -version #出现版本信息

5.安装mesos主从，看其他博客
安装完成，会在/usr/local/lib下生成一个libmesos.so文件

6.安装和配置Hadoop
master和slave

tar zxvf hadoop-2.5.0-cdh5.4.8.tar.gz
mv hadoop-2.5.0-cdh5.4.8  /usr/hadoop
cd  /usr/hadoop
mkdir -p tmp

cd /usr/hadoop/
mv bin bin-mapreduce2/
ln -s bin-mapreduce1  bin
mv example  example-mapreduce2
ln -s example-mapreduce1  example

cd etc/
mv hadoop hadoop-mapreduce2
ln -s hadoop-mapreduce1 hadoop

7.添加hadoop环境变量

vim /etc/profile
export HADOOP_HOME=/usr/hadoop
export PATH=$PATH:$HADOOP_HOME:$HADOOP_HOME/bin
source /etc/profile

8.获取hadoop on mesos的jar包

yum -y install maven openjdk-7-jdk git
git clone
cd hadoop
mvn package   #获取jar包,jar包会在target下

9.把获取的jar包放到hadoop安装目录下
master和slave

cp hadoop/target/hadoop-mesos-0.1.0.jar  /usr/hadoop/share/hadoop/common/lib/

10.配置hadoop on mesos
master和slave

vim  /usr/hadoop/etc/hadoop/mapred.site.xml
<property>
<name>mapred.job.tracker</name>
<value>localhost:9001</value>
</property>
<property>
<name>mapred.jobtracker.taskScheduler</name>
<value>org.apache.hadoop.mapred.MesosScheduler</value>
</property>
<property>
<name>mapred.mesos.taskScheduler</name>
<value>org.apache.hadoop.mapred.JobQueueTaskScheduler</value>
</property>
<property>
<name>mapred.mesos.master</name>
<value>zk://192.168.0.223</value>
</property>
<property>
<name>mapred.mesos.executor.uri</name>
<value>hdfs://localhost:9000/hadoop-2.5.0-cdh5.2.0.tar.gz</value>
</property>

11.给hadoop用户权限
master和slave

chown -R  hadoop:hadoop /usr/hadoop

12.在master上启动jobtracker，连接mesos

su hadoop
MESOS_NATIVE_LIBRARY=/usr/local/lib/libmesos.so hadoop jobtracker

13.测试

输入 192.168.0.223:5050 看看框架里有没有hadoop




----------------------------------------------------------------------------------------
二、搭建完HDFS，搭建hadoop on mesos
按照搭建hadoop文档配置完HDFS，需要配置core-site.xml，hdfs-site.xml，mapred-site.xml

mv bin  bin-mapreduce2
ln -s   bin-mapreduce1  bin
#不用移动hadoop，搭建hadoop on  mesos 的时候直接在hadoop里修改

2.移动hdfs命令和start-dfs.sh

cd /usr/hadoop/bin-mapreduce2
cp  hdfs  /usr/hadoop/bin-mapreduce1
cd /usr/hadoop/sbin
cp start-dfs.sh /sur/hadoop/bin-mapreduce1

3.搭建hadoop on mesos
修改mapred-site.xml配置文件

<property>
<name>mapred.job.tracker</name>
<value>localhost:9002</value>  #改为9002 避免和hdfs-site.xml中的端口冲突
</property>
<property>
<name>mapred.jobtracker.taskScheduler</name>
<value>org.apache.hadoop.mapred.MesosScheduler</value>
</property>
<property>
<name>mapred.mesos.taskScheduler</name>
<value>org.apache.hadoop.mapred.JobQueueTaskScheduler</value>
</property>
<property>
<name>mapred.mesos.master</name>
<value>zk://192.168.0.223</value>
</property>
<property>
<name>mapred.mesos.executor.uri</name>
<value>hdfs://localhost:9000/hadoop-2.5.0-cdh5.2.0.tar.gz</value>
</property>

4.格式化和启动HDFS

hdfs namenode  -foramt
start-dfs.sh

5.启动hadoop on mesos

su hadoop
MESOS_NATIVE_LIBRARY=/usr/local/lib/libmesos.so hadoop jobtracker

本文出自 “翟军铭的linux博客” 博客，请务必保留此出处http://zhaijunming5.blog.51cto.com/10668883/1711486