一 机器
192.168.0.203 hd203: hadoop namenode & hbase HMaster
192.168.0.204 hd204: hadoop datanode & hbase HRegionServer & zookeeper
192.168.0.205 hd205: hadoop datanode & hbase HRegionServer & zookeeper
192.168.0.206 hd206: hadoop datanode & hbase HRegionServer & zookeeper
192.168.0.202 hd202: hadoop second namenode & hbase HMaster
共 5 台机器 (生产环境需要把zookeeper单独安装)
二 操作系统和软件版本
1 操作系统:CentOS release 5.6 (Final)x64,关闭防火墙
2 相关软件:
jdk-6u24-linux-x64.bin
hadoop-0.20.2-cdh3u0.tar.gz
hbase-0.90.1-cdh3u0.tar.gz
zookeeper-3.3.3-cdh3u0.tar.gz
三 安装
1 安装jdk (所有机器)
mkdir /usr/java
mv jdk-6u24-linux-x64.bin /usr/java
chmod 744 jdk-6u24-linux-x64.bin
./jdk-6u24-linux-x64.bin
ln -s jdk1.6.0_24 default
2 建立集群平台用户 (所有机器)
useradd cbcloud
passwd cbcloud 密码设为123456
3 编辑/etc/hosts (所有机器)
192.168.0.203 hd203
192.168.0.204 hd204
192.168.0.205 hd205
192.168.0.206 hd206
192.168.0.202 hd202
4 hd203登录集群所有机器免密码设置 (hd203)
su - cbcloud
ssh登录本机也要免密码 如下设置
$ ssh-keygen -t dsa -P '' -f ~/.ssh/id_dsa
$ cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys
将hd203上的id_dsa.pub 文件追加到所有机器的cbcloud用户的authorized_keys文件内
细节不在详述
ssh 所有机器 第一次需要确认一下
5 安装hadoop
5.1 建立目录 (所有机器)
mkdir /home/cbcloud/hdtmp
mkdir /home/cbcloud/hddata
mkdir /home/cbcloud/hdconf
chown -R cbcloud:cbcloud 以上目录
tar zxvf hadoop-0.20.2-cdh3u0.tar.gz -C /home/cbcloud
cd /home/cbcloud
mv hadoop-0.20.2-cdh3u0 hadoop
chown -R cbcloud:cbcloud hadoop/
5.2 配置环境变量 vi /etc/profile (所有机器)
export JAVA_HOME=/usr/java/default
export CLASSPATH=$CLASSPATH:$JAVA_HOME/lib:$JAVA_HOME/jre/lib
export PATH=$JAVA_HOME/bin:$JAVA_HOME/lib:$JAVA_HOME/jre/bin:$PATH:$HOME/bin
export HADOOP_HOME=/home/cbcloud/hadoop
export HADOOP_CONF_DIR=/home/cbcloud/hdconf
export PATH=$PATH:$HADOOP_HOME/bin
把hadoop的配置文件目录从源目录拿出来,方便以后升级hadoop
mv hadoop的conf目录内文件到/home/cbcloud/hdconf内
5.3 编辑 hadoop 配置文件 core-site.xml
加入
<configuration>
<property>
<name>fs.default.name</name>
<value>hdfs://hd203:9000</value>
</property>
<property>
<name>fs.checkpoint.dir</name>
<value>/home/cbcloud/hdtmp/dfs/namesecondary</value>
<description>Determines where on the local filesystem the DFS secondary
name node should store the temporary images to merge.
If this is a comma-delimited list of directories then the image is
replicated in all of the directories for redundancy.
</description>
</property>
<property>
<name>fs.checkpoint.period</name>
<value>60</value>
<description>Determines where on the local filesystem the DFS secondary
name node should store the temporary images to merge.
If this is a comma-delimited list of directories then the image is
replicated in all of the directories for redundancy.
</description>
</property>
5.4 编辑hdfs-site.xml
加入
<property>
<name>dfs.replication</name>
<value>3</value>
</property>
<property>
<name>dfs.data.dir</name>
<value>/home/cbcloud/hddata</value>
</property>
<property>
<name>hadoop.tmp.dir</name>
<value>/home/cbcloud/hdtmp/</value>
</property>
<property>
<name>dfs.balance.bandwidthPerSec</name>
<value>10485760</value>
</property>
<property>
<name>dfs.hosts.exclude</name>
<value>/home/cbcloud/hdconf/excludes</value>
<final>true</final>
</property>
5.5 编辑mapred-site.xml
加入
<property>
<name>mapred.job.tracker</name>
<value>hd203:9001</value>
</property>
5.6 编辑 hadoop-env.sh
export JAVA_HOME=/usr/java/default
5.7 编辑masters 该文件指定 secondary name 机器,
加入
hd202
编辑 slaves
加入
hd204
hd205
hd206
5.8 拷贝hd203的hadoop和hdconf到所有机器
# scp -r /home/cbcloud/hadoop cbcloud@hd204:/home/cbcloud
# scp -r /home/cbcloud/hdconf cbcloud@hd204:/home/cbcloud
完成后,在hd203 格式化hadoop文件系统
执行
hadoop namenode -format
启动
start-all.sh
查看集群内datanode的机器
执行jps
5764 Jps
18142 DataNode
18290 TaskTracker
看到以上结果 说明启动正确
web方式
http://hd203:50070/dfshealth.jsp
注意 本地PC hosts文件也要配置
192.168.0.203 hd203
192.168.0.204 hd204
192.168.0.205 hd205
192.168.0.206 hd206
192.168.0.202 hd202
web方式可以查看集群状态和job状态等,至此hadoop安装完毕
6 安装zookeeper (hd203)
tar zxvf zookeeper-3.3.3-cdh3u0.tar.gz -C /home/cbcloud
在hd204-hd206上
mkdir /home/cbcloud/zookeeperdata
chown -R cbcloud:cbcloud /home/cbcloud/zookeeperdata
chown -R cbcloud:cbcloud /home/cbcloud/zookeeper-3.3.3-cdh3u0
编辑 /home/cbcloud/zookeeper-3.3.3-cdh3u0/conf/zoo.cfg
# The number of milliseconds of each tick
tickTime=2000
# The number of ticks that the initial
# synchronization phase can take
initLimit=10
# The number of ticks that can pass between
# sending a request and getting an acknowledgement
syncLimit=5
# the directory where the snapshot is stored.
dataDir=/home/cbcloud/zookeeperdata
# the port at which the clients will connect
clientPort=2181
server.1=hd204:2888:3888
server.2=hd205:2888:3888
server.3=hd206:2888:3888
scp hd203的zookeeper到hd204,hd205,hd206
# scp -r /home/cbcloud/zookeeper-3.3.3-cdh3u0/ cbcloud@hd205:/home/cbcloud/
在hd204-206 的/home/cbcloud/zookeeperdata目录touch myid文件,
内容分别为1,2,3 和server编号一致 chown cbcloud:cbcloud myid
启动zookeeper,在hd204-206上bin目录下 执行
# zkServer.sh start
启动后 通过
# zkServer.sh status
查看状态 注意 在centos5.6上 执行这个会报错
Error contacting service. It is probably not running.
通过查看脚本 是因为
echo stat | nc -q 1 localhost
nc版本不同,没有-q的参数,更改脚本去掉-q 1 即可
另外 可以通过
echo stat | nc localhost 2181来查看状态
7 安装hbase
7.1 建立目录 (所有机器)
mkdir /home/cbcloud/hbconf
chown -R cbcloud:cbcloud /home/cbcloud/hbconf
tar zxvf hbase-0.90.1-cdh3u0.tar.gz -C /home/cbcloud
cd /home/cbcloud
mv hbase-0.90.1-cdh3u0 hbase
chown -R cbcloud:cbcloud hbase/
7.2 配置环境变量
vi /etc/profile (所有机器) 追加如下内容
export HBASE_CONF_DIR=/home/cbcloud/hbconf
export HBASE_HOME=/home/hadoop/hbase
把hbase的配置文件目录从源目录拿出来,方便以后升级hbase
mv hbase的conf目录内文件到/home/cbcloud/hbconf内
7.3 编辑 hbase-env.sh
export HBASE_OPTS="$HBASE_OPTS -XX:+HeapDumpOnOutOfMemoryError -XX:+UseConcMarkSweepGC -XX:+CMSIncrementalMode"
export JAVA_HOME=/usr/java/default
export HBASE_MANAGES_ZK=false
export HBASE_HOME=/home/cbcloud/hbase
export HADOOP_HOME=/home/cbcloud/hadoop
7.4 编辑hbase-site.xml
加入
<property>
<name>hbase.rootdir</name>
<value>hdfs://hd203:9000/hbase</value>
</property>
<property>
<name>hbase.cluster.distributed</name>
<value>true</value>
</property>
<property>
<name>hbase.master</name>
<value>hd203:60000</value>
</property>
<property>
<name>hbase.master.port</name>
<value>60000</value>
<description>The port master should bind to.</description>
</property>
<property>
<name>hbase.zookeeper.quorum</name>
<value>hd204,hd205,hd206</value>
</property>
7.5 编辑regionservers
加入
hd204
hd205
hd206
scp hd203 的hbase到hd204-206,202
# scp -r /home/cbcloud/hbase/ cbcloud@hd204:/home/cbcloud
# scp -r /home/cbcloud/hbconf/ cbcloud@hd204:/home/cbcloud
su - cbcloud
启动hbase
在hd203上hbase的bin目录执行
$ ./start-hbase.sh
启动hbase的第二个HMaster
在202上执行
$ ./hbase-daemon.sh start master
查看Master:http://hd203:60010/master.jsp
此时查看datanode 服务器 jps
5764 Jps
32091 HRegionServer
18142 DataNode
11221 QuorumPeerMain
18290 TaskTracker
以上结果说明启动正常
至此 hadoop+zookeeper+hbase安装完成
启动顺序
1.hadoop
2.zookeeper
3.hbase
4.第二个HMaster
停止顺序
1.第二个HMaster, kill-9删除
2.hbase
3.zookeeper
4.hadoop
注意 一定要按顺序停止,
如果先停zookeeper再停hbase的话,基本停不下来(自己测试结果)