基于docker1.7.03.1单机上部署hadoop2.7.3分布式集群

[TOC]

声明

文章均为本人技术笔记，转载请注明出处：
[1] https://segmentfault.com/u/yzwall
[2] blog.csdn.net/j_dark/

0 docker版本与hadoop版本说明

• PC：ubuntu 16.04.1 LTS

• Docker version：17.03.1-ce OS/Arch:linux/amd64

• Hadoop version：hadoop-2.7.3

1 docker中配置构建hadoop镜像

1.1 创建docker容器container

创建基于ubuntu镜像的容器container，官方默认下载ubuntu最新精简版镜像；
sudo docker run -ti container ubuntu

1.2 修改/etc/source.list

修改默认源文件/etc/apt/source.list，用国内源代替官方源；

1.3 安装java8

# docker镜像为了精简容量，删除了许多ubuntu自带组件，通过`apt-get update`更新获得
apt-get update
apt-get install software-properties-common python-software-properties # add-apt-repository
apt-get install software-properties-commonapt-get install software-properties-common # add-apt-repository
add-apt-repository ppa:webupd8team/java
apt-get update
apt-get install oracle-java8-installer
java -version

1.4 docker中安装hadoop-2.7.3

1.4.1 下载hadoop-2.7.3源码

# 创建多级目录
mkdir -p /software/apache/hadoop
cd /software/apache/hadoop
# 下载并解压hadoop
wget http://mirrors.sonic.net/apache/hadoop/common/hadoop-2.7.3/hadoop-2.7.3.tar.gz
tar xvzf hadoop-2.7.3.tar.gz

1.4.2 配置环境变量

修改~/.bashrc文件。在文件末尾加入下面配置信息：

export JAVA_HOME=/usr/lib/jvm/java-8-oracle
export HADOOP_HOME=/software/apache/hadoop/hadoop-2.7.3
export HADOOP_CONFIG_HOME=$HADOOP_HOME/etc/hadoop
export PATH=$PATH:$HADOOP_HOME/bin
export PATH=$PATH:$HADOOP_HOME/sbin

source ~/.bashrc使环境变量配置生效；
注意：完成./bashrc文件配置后，hadoop-env.sh无需再配置；

1.5 配置hadoop

配置hadoop主要配置core-site.xml、hdfs-site.xml、mapred-site.xml， yarn-site.xml三个文件；

在$HADOOP_HOME下创建namenode, datanode和tmp目录

cd $HADOOP_HOME
mkdir tmp
mkdir namenode
mkdir datanode

1.5.1 配置core.site.xml

• 配置项hadoop.tmp.dir指向tmp目录

• 配置项fs.default.name指向master节点，配置为hdfs://master:9000

<configuration><property><!-- hadoop temp dir  --><name>hadoop.tmp.dir</name><value>/software/apache/hadoop/hadoop-2.7.3/tmp</value><description>A base for other temporary directories.</description></property><!-- Size of read/write buffer used in SequenceFiles. --><property><name>io.file.buffer.size</name><value>131072</value></property><property><name>fs.default.name</name><value>hdfs://master:9000</value><final>true</final><description>The name of the default file system.</description></property>
</configuration>

1.5.2 配置hdfs-site.xml

• dfs.replication表示节点数目，配置集群1个namenode，3个datanode，设置备份数为4；

• dfs.namenode.name.dir和dfs.datanode.data.dir分别配置为之前创建的NameNode和DataNode的目录路径

<configuration><property><name>dfs.namenode.secondary.http-address</name><value>master:9001</value></property><property><name>dfs.replication</name><value>3</value><final>true</final><description>Default block replication.</description></property><property><name>dfs.namenode.name.dir</name><value>/software/apache/hadoop/hadoop-2.7.3/namenode</value><final>true</final></property><property><name>dfs.datanode.data.dir</name><value>/software/apache/hadoop/hadoop-2.7.3/datanode</value><final>true</final></property><property><name>dfs.webhdfs.enabled</name><value>true</value></property>
</configuration>

1.5.3 配置mapred-site.xml

在$HADOOP_HOME下使用cp命令创建mapred-site.xml

cd $HADOOP_HOME
cp mapred-site.xml.template mapred-site.xml

配置mapred-site.xml，配置项mapred.job.tracker指向master节点；

在hadoop 2.x.x中，用户无需配置mapred.job.tracker，因为JobTracker已经不存在，功能由组件MRAppMaster实现，因此需要用mapreduce.framework.name指定运行框架名称，指定yarn

——《Hadoop技术内幕：深入解析YARN架构设计与实现原理》

<configuration><property><name>mapreduce.framework.name</name><value>yarn</value></property><property><name>mapreduce.jobhistory.address</name><value>master:10020</value></property><property><name>mapreduce.jobhistory.address</name><value>master:19888</value></property>
</configuration>

1.5.4 配置yarn-site.xml

<configuration><property>  <name>yarn.nodemanager.aux-services</name>  <value>mapreduce_shuffle</value>  </property>  <property>                                                                  <name>yarn.nodemanager.aux-services.mapreduce.shuffle.class</name>  <value>org.apache.hadoop.mapred.ShuffleHandler</value>  </property>  <property>  <name>yarn.resourcemanager.address</name>  <value>master:8032</value>  </property>  <property>  <name>yarn.resourcemanager.scheduler.address</name>  <value>master:8030</value>  </property>  <property>  <name>yarn.resourcemanager.resource-tracker.address</name>  <value>master:8031</value>  </property>  <property>  <name>yarn.resourcemanager.admin.address</name>  <value>master:8033</value>  </property>  <property>  <name>yarn.resourcemanager.webapp.address</name>  <value>master:8088</value>  </property>
</configuration>

1.5.5 安装vim,ifconfig与ping

安装ifconfig与ping命令所需软件包

apt-get update
apt-get install vim
apt-get install net-tools       # for ifconfig
apt-get install inetutils-ping  # for ping

1.5.6 构建hadoop基础镜像

假设当前容器名为container，保存基础镜像为ubuntu:hadoop，后续hadoop集群容器都根据该镜像创建启动，无需重复配置；
sudo docker commit -m "hadoop installed" container ubuntu:hadoop /bin/bash

2. hadoop分布式集群搭建

2.1 根据已经创建hadoop基础镜像创建容器集群

分别根据基础镜像ubuntu:hadoop创建mater容器和slave1~3容器，各自主机名与容器名一致；
创建master：docker run -ti -h master --name master ubuntu:hadoop /bin/bash
创建slave1：docker run -ti -h slave1 --name slave1 ubuntu:hadoop /bin/bash
创建slave2：docker run -ti -h slave2 --name slave2 ubuntu:hadoop /bin/bash
创建slave3：docker run -ti -h slave3 --name slave3 ubuntu:hadoop /bin/bash

2.2 配置各容器hosts文件

在各容器的/etc/hosts中添加以下内容，各容器ip地址通过ifconfig查看：

master 172.17.0.2
slave1 172.17.0.3
slave2 172.17.0.4
slave3 172.17.0.5 

注意：docker容器重启后，hosts内容可能会失效，经验不足暂时只能避免容器频繁重启，否则得手动再次配置hosts文件；

参考http://dockone.io/question/400

1./etc/hosts, /etc/resolv.conf和/etc/hostname，容器中的这三个文件不存在于镜像，而是存在于/var/lib/docker/containers/<container_id>，在启动容器的时候，通过mount的形式将这些文件挂载到容器内部。因此，如果在容器中修改这些文件的话，修改部分不会存在于容器的top layer，而是直接写入这三个物理文件中。
2.为什么重启后修改内容不存在？原因是：每次Docker在启动容器的时候，通过重新构建新的/etc/hosts文件，这又是为什么呢？原因是：容器重启，IP地址为改变，hosts文件中原来的IP地址无效，因此理应修改hosts文件，否则会产生脏数据。？原因是：每次Docker在启动容器的时候，通过重新构建新的/etc/hosts文件，这又是为什么呢？原因是：容器重启，IP地址为改变，hosts文件中原来的IP地址无效，因此理应修改hosts文件，否则会产生脏数据。1./etc/hosts, /etc/resolv.conf和/etc/hostname，容器中的这三个文件不存在于镜像，而是存在于/var/lib/docker/containers/<container_id>，在启动容器的时候，通过mount的形式将这些文件挂载到容器内部。因此，如果在容器中修改这些文件的话，修改部分不会存在于容器的top layer，而是直接写入这三个物理文件中。

2.3 集群节点SSH配置

2.3.1 所有节点：安装ssh

apt-get update
apt-get install ssh
apt-get install openssh-server

2.3.2 所有节点：生成随机密钥

# 生成无密码密钥，生成密钥位于~/.ssh下
ssh-keygen -t rsa -P ""

2.3.3 master节点：生成证书文件authorized_keys

将生成的公钥写入authorized_keys中

cat ~/.ssh/id_rsa.pub >> ~/.ssh/authorized_keys 

2.3.4 所有节点：修改sshd_config文件

通过修改sshd_config文件，保证ssh可远程登陆其他节点的root用户

vim /etc/ssh/sshd_config
# 将PermitRootLogin prohibit-password修改为PermitRootLogin yes
# 重启ssh服务
service ssh restart

2.3.5 master节点：通过scp传输证书到slave节点

传输master节点上的authorized_keys到其他slave节点~/.ssh下，覆盖同名文件；保证所有节点的证书一致，因此可以实现任意节点间可以通过ssh访问；

cd ~/.ssh
scp authorized_keys root@slave1:~/.ssh/
scp authorized_keys root@slave2:~/.ssh/
scp authorized_keys root@slave3:~/.ssh/

2.3.6 slave节点：修改证书权限确保生效

chmod 600 ~/.ssh/authorized_keys

注意

• 查看ssh服务是否开启：ps -e | grep ssh

• 开启ssh服务：service ssh start

• 重启ssh服务：service ssh restart

完成2.3.1操作后，各个容器之间可通过ssh访问；

2.4 master节点配置

在master节点中，修改slaves文件配置slave节点

cd $HADOOP_CONFIG_HOME/
vim slaves

将其中内容覆盖为：

slave1
slave2
slave3

2.5 启动hadoop集群

进入master节点，

• 执行hdfs namenode -format，出现类似信息表示namenode格式化成功：

common.Storage: Storage directory /software/apache/hadoop/hadoop-2.7.3/namenode has been successfully formatted.

• 执行start_all.sh启动集群：

root@master:/# start-all.sh
This script is Deprecated. Instead use start-dfs.sh and start-yarn.sh
Starting namenodes on [master]
The authenticity of host 'master (172.17.0.2)' can't be established.
ECDSA key fingerprint is SHA256:OewrSOYpvfDE6ixf6Gw9U7I9URT2zDCCtDJ6tjuZz/4.
Are you sure you want to continue connecting (yes/no)? yes
master: Warning: Permanently added 'master,172.17.0.2' (ECDSA) to the list of known hosts.
master: starting namenode, logging to /software/apache/hadoop/hadoop-2.7.3/logs/hadoop-root-namenode-master.out
slave3: starting datanode, logging to /software/apache/hadoop/hadoop-2.7.3/logs/hadoop-root-datanode-slave3.out
slave2: starting datanode, logging to /software/apache/hadoop/hadoop-2.7.3/logs/hadoop-root-datanode-slave2.out
slave1: starting datanode, logging to /software/apache/hadoop/hadoop-2.7.3/logs/hadoop-root-datanode-slave1.out
Starting secondary namenodes [master]
master: starting secondarynamenode, logging to /software/apache/hadoop/hadoop-2.7.3/logs/hadoop-root-secondarynamenode-master.out
starting yarn daemons
starting resourcemanager, logging to /software/apache/hadoop/hadoop-2.7.3/logs/yarn-root-resourcemanager-master.out
slave3: starting nodemanager, logging to /software/apache/hadoop/hadoop-2.7.3/logs/yarn-root-nodemanager-slave3.out
slave1: starting nodemanager, logging to /software/apache/hadoop/hadoop-2.7.3/logs/yarn-root-nodemanager-slave1.out
slave2: starting nodemanager, logging to /software/apache/hadoop/hadoop-2.7.3/logs/yarn-root-nodemanager-slave2.out

分别在master，slave节点中执行jps，

• master:

root@master:/# jps
2065 Jps
1446 NameNode
1801 ResourceManager
1641 SecondaryNameNode

• slave1:

1107 NodeManager
1220 Jps
1000 DataNode

• slave2:

241 DataNode
475 Jps
348 NodeManager

• slave3:

500 Jps
388 NodeManager
281 DataNode

3. 执行wordcount

在hdfs中创建输入目录/hadoopinput，并将输入文件LICENSE.txt存储在该目录下：

root@master:/# hdfs dfs -mkdir -p /hadoopinput
root@master:/# hdfs dfs -put LICENSE.txt /hadoopint

进入$HADOOP_HOME/share/hadoop/mapreduce，提交wordcount任务给集群，将计算结果保存在hdfs中的/hadoopoutput目录下：

root@master:/# cd $HADOOP_HOME/share/hadoop/mapreduce
root@master:/software/apache/hadoop/hadoop-2.7.3/share/hadoop/mapreduce# hadoop jar hadoop-mapreduce-examples-2.7.3.jar wordcount /hadoopinput /hadoopoutput
17/05/26 01:21:34 INFO client.RMProxy: Connecting to ResourceManager at master/172.17.0.2:8032
17/05/26 01:21:35 INFO input.FileInputFormat: Total input paths to process : 1
17/05/26 01:21:35 INFO mapreduce.JobSubmitter: number of splits:1
17/05/26 01:21:35 INFO mapreduce.JobSubmitter: Submitting tokens for job: job_1495722519742_0001
17/05/26 01:21:36 INFO impl.YarnClientImpl: Submitted application application_1495722519742_0001
17/05/26 01:21:36 INFO mapreduce.Job: The url to track the job: http://master:8088/proxy/application_1495722519742_0001/
17/05/26 01:21:36 INFO mapreduce.Job: Running job: job_1495722519742_0001
17/05/26 01:21:43 INFO mapreduce.Job: Job job_1495722519742_0001 running in uber mode : false
17/05/26 01:21:43 INFO mapreduce.Job:  map 0% reduce 0%
17/05/26 01:21:48 INFO mapreduce.Job:  map 100% reduce 0%
17/05/26 01:21:54 INFO mapreduce.Job:  map 100% reduce 100%
17/05/26 01:21:55 INFO mapreduce.Job: Job job_1495722519742_0001 completed successfully
17/05/26 01:21:55 INFO mapreduce.Job: Counters: 49File System CountersFILE: Number of bytes read=29366FILE: Number of bytes written=295977FILE: Number of read operations=0FILE: Number of large read operations=0FILE: Number of write operations=0HDFS: Number of bytes read=84961HDFS: Number of bytes written=22002HDFS: Number of read operations=6HDFS: Number of large read operations=0HDFS: Number of write operations=2Job Counters Launched map tasks=1Launched reduce tasks=1Data-local map tasks=1Total time spent by all maps in occupied slots (ms)=2922Total time spent by all reduces in occupied slots (ms)=3148Total time spent by all map tasks (ms)=2922Total time spent by all reduce tasks (ms)=3148Total vcore-milliseconds taken by all map tasks=2922Total vcore-milliseconds taken by all reduce tasks=3148Total megabyte-milliseconds taken by all map tasks=2992128Total megabyte-milliseconds taken by all reduce tasks=3223552Map-Reduce FrameworkMap input records=1562Map output records=12371Map output bytes=132735Map output materialized bytes=29366Input split bytes=107Combine input records=12371Combine output records=1906Reduce input groups=1906Reduce shuffle bytes=29366Reduce input records=1906Reduce output records=1906Spilled Records=3812Shuffled Maps =1Failed Shuffles=0Merged Map outputs=1GC time elapsed (ms)=78CPU time spent (ms)=1620Physical memory (bytes) snapshot=451264512Virtual memory (bytes) snapshot=3915927552Total committed heap usage (bytes)=348127232Shuffle ErrorsBAD_ID=0CONNECTION=0IO_ERROR=0WRONG_LENGTH=0WRONG_MAP=0WRONG_REDUCE=0File Input Format Counters Bytes Read=84854File Output Format Counters Bytes Written=22002

计算结果保存在/hadoopoutput/part-r-00000中，查看结果：

root@master:/# hdfs dfs -ls /hadoopoutput
Found 2 items
-rw-r--r--   3 root supergroup          0 2017-05-26 01:21 /hadoopoutput/_SUCCESS
-rw-r--r--   3 root supergroup      22002 2017-05-26 01:21 /hadoopoutput/part-r-00000root@master:/# hdfs dfs -cat /hadoopoutput/part-r-00000
""AS    2
"AS    16
"COPYRIGHTS    1
"Contribution"    2
"Contributor"    2
"Derivative    1
"Legal    1
"License"    1
"License");    1
"Licensed    1
"Licensor"    1
...

至此，基于docker1.7.03单机上部署hadoop2.7.3集群圆满成功！

参考

[1] http://tashan10.com/yong-dockerda-jian-hadoopwei-fen-bu-shi-ji-qun/
[2] http://blog.csdn.net/xiaoxiangzi222/article/details/52757168