hadoop cdh5安装

为满足公司程序运行环境，hadoop集群由原来的1.0版本升级到CDH5版本，又一次集群安装经历，分享给有需要的人。

一、机器准备

    Linux版本CentOs 5.8，x86_64，如果你的linux版本是6.x，也可以参照下面步骤安装；
本人此次安装共准备了5台机器：
192.168.32.70（master），192.168.32.71（slave1），192.168.32.72（slave2），192.168.32.73(slave3)，192.168.32.79（slave4）；
修改/etc/sysconfig/network文件中的HOSTNAME，修改为方便记忆的名字，当然你也可以不改，只要你觉得方便就好；
修改/etc/hosts文件（五台机器都要修改）：
192.168.32.70 master
192.168.32.71 slave1
192.168.32.72 slave2
192.168.32.73 slave3
192.168.32.79 slave4

二、环境准备

1、打通ssh
>所有机器 ssh-keygen -t rsa 一路按回车；
>在master机器上执行：cat ~/.ssh/id_rsa.pub >> ~/.ssh/authorized_keys；
>scp文件到其他几台机器：
    scp ~/.ssh/authorized_keys root@slave1:~/.ssh/
    scp ~/.ssh/authorized_keys root@slave2:~/.ssh/
                  ......
    scp ~/.ssh/authorized_keys root@slave4:~/.ssh/
>试验下免密码功能是否正常：

[root@master hadoop-conf]# ssh slave1
Last login: Wed Sep 24 16:07:12 2014 from master
[root@slave1 ~]#

没有提示输入密码，表示成功了；
2、安装JDK7
>官网下载jdk-7u51-linux-x64.rpm包；
>rpm -ivh jdk-7u51-linux-x64.rpm
>添加环境变量;
vi /etc/profile
增加
     JAVA_HOME=/usr/java/latest
     PATH=$PATH:$JAVA_HOME/bin
     CLASSPATH=$JAVA_HOME/lib:$JAVA_HOME/jre/lib
     export JAVA_HOME CLASSPATH
>执行source生效;
source /etc/profile
3、创建hadoop用户
    groupadd hdfs
    useradd hadoop -g hdfs

三、安装cdh5

1、下载rpm安装包
>进入目录/data/tools/ (个人习惯的软件存储目录，你可以自己随便选择)；
    wget "5/x86_64/cloudera-cdh-5-0.x86_64.rpm"   ---------如果你的Linux版本是6.x这里改为6即可，下同；
    yum --nogpgcheck localinstall cloudera-cdh-5-0.x86_64.rpm
>添加cloudera仓库验证;
    rpm --import 5/x86_64/cdh/RPM-GPG-KEY-cloudera
2、安装
> master 安装NN，NM，DN，MR，hadoop-client
yum clean all; yum install hadoop-hdfs-namenode
yum clean all; yum install hadoop-yarn-nodemanager hadoop-hdfs-datanode hadoop-mapreduce
yum clean all; yum install hadoop-client
>slave1 安装RM，NM，DN，MR，hadoop-client
yum clean all; yum install hadoop-yarn-resourcemanager
yum clean all; yum install hadoop-yarn-nodemanager hadoop-hdfs-datanode hadoop-mapreduce
yum clean all; yum install hadoop-client
>slave2 、slave3、slave4安装NM，DN，MR，hadoop-client
yum clean all; yum install hadoop-yarn-nodemanager hadoop-hdfs-datanode hadoop-mapreduce
yum clean all; yum install hadoop-client
3、创建目录 (本人机器只有一个盘cache1，如果你有多个可以创建多个)
DN：
mkdir -p /data/cache1/dfs/dn
mkdir -p /data/cache1/dfs/mapred/local
chown -R hdfs:hadoop /data/cache1/dfs/dn
chown -R mapred:hadoop /data/cache1/dfs/mapred/local
NN：
mkdir -p /data/cache1/dfs/nn
chown -R hdfs:hadoop /data/cache1/dfs/nn
chmod 700 /data/cache1/dfs/nn
4、修改配置文件
修改master机器上的配置文件，然后scp到各个slave；
1)/etc/hadoop/conf/core-site.xml   红色IP为NN地址；

[root@master conf]# cat core-site.xml






fs.defaultFS
hdfs://192.168.32.70:8020



dfs.replication
1

2)/etc/hadoop/conf/hdfs-site.xml/yarn-site.xml

[root@master conf]# cat /etc/hadoop/conf/hdfs-site.xml






dfs.name.dir
/var/lib/hadoop-hdfs/cache/hdfs/dfs/name


dfs.datanode.data.dir
/data/cache1/dfs/dn/

3)/etc/hadoop/conf  红色的IP为装RM的机器，本例子中是192.168.32.71；

[root@master conf]# cat yarn-site.xml






yarn.nodemanager.aux-services
mapreduce_shuffle



yarn.nodemanager.aux-services.mapreduce_shuffle.class
org.apache.hadoop.mapred.ShuffleHandler



yarn.log-aggregation-enable
true



List of directories to store localized files in.
yarn.nodemanager.local-dirs
/var/lib/hadoop-yarn/cache/${user.name}/nm-local-dir


yarn.resourcemanager.address
192.168.32.71:8032



yarn.resourcemanager.scheduler.address
192.168.32.71:8030



yarn.resourcemanager.webapp.address
0.0.0.0:8088



yarn.resourcemanager.resource-tracker.address
192.168.32.71:8031



yarn.resourcemanager.admin.address
192.168.32.71:8033






Where to store container logs.
yarn.nodemanager.log-dirs
/var/log/hadoop-yarn/containers


Where to aggregate logs to.
yarn.nodemanager.remote-app-log-dir
/var/log/hadoop-yarn/apps



Classpath for typical applications.
yarn.application.classpath

$HADOOP_CONF_DIR,
$HADOOP_COMMON_HOME/*,$HADOOP_COMMON_HOME/lib/*,
$HADOOP_HDFS_HOME/*,$HADOOP_HDFS_HOME/lib/*,
$HADOOP_MAPRED_HOME/*,$HADOOP_MAPRED_HOME/lib/*,
$HADOOP_YARN_HOME/*,$HADOOP_YARN_HOME/lib/*

4)/etc/hadoop/conf/hadoop-env.sh 

[root@master conf]# cat hadoop-env.sh
# Set Hadoop-specific environment variables here.

# The only required environment variable is JAVA_HOME.  All others are
# optional.  When running a distributed configuration it is best to
# set JAVA_HOME in this file, so that it is correctly defined on
# remote nodes.

# The maximum amount of heap to use, in MB. Default is 1000.
#export HADOOP_HEAPSIZE=
#export HADOOP_NAMENODE_INIT_HEAPSIZE=""

# Extra Java runtime options.  Empty by default.
export HADOOP_OPTS="-Djava.net.preferIPv4Stack=true ${HADOOP_OPTS}"

# Command specific options appended to HADOOP_OPTS when specified
export HADOOP_NAMENODE_OPTS="-Dsecurity.audit.logger=INFO,DRFAS -Dhdfs.audit.logger=INFO,DRFAAUDIT ${HADOOP_NAMENODE_OPTS}"
HADOOP_JOBTRACKER_OPTS="-Dsecurity.audit.logger=INFO,DRFAS -Dmapred.audit.logger=INFO,MRAUDIT -Dmapred.jobsummary.logger=INFO,JSA ${HADOOP_JOBTRACKER_OPTS}"
HADOOP_TASKTRACKER_OPTS="-Dsecurity.audit.logger=ERROR,console -Dmapred.audit.logger=ERROR,console ${HADOOP_TASKTRACKER_OPTS}"
HADOOP_DATANODE_OPTS="-Dsecurity.audit.logger=ERROR,DRFAS ${HADOOP_DATANODE_OPTS}"

export HADOOP_SECONDARYNAMENODE_OPTS="-Dsecurity.audit.logger=INFO,DRFAS -Dhdfs.audit.logger=INFO,DRFAAUDIT ${HADOOP_SECONDARYNAMENODE_OPTS}"

# The following applies to multiple commands (fs, dfs, fsck, distcp etc)
export HADOOP_CLIENT_OPTS="-Xmx128m ${HADOOP_CLIENT_OPTS}"
#HADOOP_JAVA_PLATFORM_OPTS="-XX:-UsePerfData ${HADOOP_JAVA_PLATFORM_OPTS}"

# On secure datanodes, user to run the datanode as after dropping privileges
export HADOOP_SECURE_DN_USER=hdfs

# Where log files are stored.  $HADOOP_HOME/logs by default.
export HADOOP_LOG_DIR=/var/local/hadoop/logs

# Where log files are stored in the secure data environment.
export HADOOP_SECURE_DN_LOG_DIR=$HADOOP_LOG_DIR

# The directory where pid files are stored. /tmp by default.
export HADOOP_PID_DIR=/var/local/hadoop/pid
export HADOOP_SECURE_DN_PID_DIR=$HADOOP_PID_DIR

# A string representing this instance of hadoop. $USER by default.
export HADOOP_IDENT_STRING=$USER
export JAVA_HOME=/usr/java/latest

5)修改/etc/hadoop/conf/slave文件；
添加slave：
slave1
slave2
slave3
slave4
6)scp文件到各个slave；
scp /etc/hadoop/conf root@slave1:/etc/hadoop/conf

四、启动

1)NN(master)启动
/etc/init.d/hadoop-hdfs-namenode init
/etc/init.d/hadoop-hdfs-namenode start
2)DN(slave1)启动(装有RM）
/etc/init.d/hadoop-datanode
/etc/init.d/hadoop-yarn-nodemanager
/etc/init.d/hadoop-yarn-resourcemanager
3)DN(slave2/slave3/slave4)启动
/etc/init.d/hadoop-datanode
/etc/init.d/hadoop-yarn-nodemanager

五、查看

   
  （类似于hadoop1.0的Jobtracker地址，即50030端口）

六、安装中出现的问题以及解决办法

启动NN时报：log4j:ERROR Could not find value for key log4j.appender.DRFAAUDIT错误；
解决办法：在/etc/hadoop/conf/log4j.properties 加入以下配置
log4j.appender.DRFAAUDIT=org.apache.log4j.ConsoleAppender
log4j.appender.DRFAAUDIT.layout=org.apache.log4j.PatternLayout

七、总结

hadoop 1.0版本和当前装的cdh5版本，从安装方面差别还是挺大的，不过还好不算麻烦，一步一步来，遇到问题多问就OK；