Hadoop安装杂记（2）

一、分布式模型

1、环境准备

准备4个节点，master1为主控节点（NameNode、SecondaryNameNode、ResourceManager），master2-4作为数据节点（DataNode、NodeManager）。并做好ntp时间同步

1.1 每个节点配置JAVA环境

[root@master1 ~]# vim /etc/profile.d/java.sh

export JAVA_HOME=/usr

[root@master1 ~]# scp /etc/profile.d/java.sh root@master2:/etc/profile.d/
[root@master1 ~]# scp /etc/profile.d/java.sh root@master3:/etc/profile.d/
[root@master1 ~]# scp /etc/profile.d/java.sh root@master4:/etc/profile.d/


每个节点安装java-devel
[root@master1 ~]# yum install -y java-1.7.0-openjdk-devel
[root@master2 ~]# yum install -y java-1.7.0-openjdk-devel
[root@master3 ~]# yum install -y java-1.7.0-openjdk-devel
[root@master4 ~]# yum install -y java-1.7.0-openjdk-devel

配置hadoop环境变量：
[root@master1 ~]# vim /etc/profile.d/hadoop.sh

export HADOOP_PREFIX=/bdapps/hadoop
export PATH=$PATH:${HADOOP_PREFIX}/bin:${HADOOP_PREFIX}/sbin
export HADOOP_YARN_HOME=${HADOOP_PREFIX}
export HADOOP_MAPPERD_HOME=${HADOOP_PREFIX}
export HADOOP_COMMON_HOME=${HADOOP_PREFIX}
export HADOOP_HDFS_HOME=${HADOOP_PREFIX}

[root@master1 ~]# source /etc/profile.d/hadoop.sh

scp /etc/profile.d/hadoop.sh master2:/etc/profile.d/hadoop.sh
scp /etc/profile.d/hadoop.sh master3:/etc/profile.d/hadoop.sh
scp /etc/profile.d/hadoop.sh master4:/etc/profile.d/hadoop.sh

1.2 每个节点准备host文件，实验使用别名调用

[root@master1 ~]# cat /etc/hosts
127.0.0.1   localhost localhost.localdomain localhost4 localhost4.localdomain4
::1         localhost localhost.localdomain localhost6 localhost6.localdomain6
10.201.106.131 master1 master1.com master
10.201.106.132 master2 master2.com
10.201.106.133 master3 master3.com
10.201.106.134 master4 master4.com

master2,3,4节点同上

1.3 创建用户组和用户

[root@master1 ~]# useradd -g hadoop hadoop

设置用户密码：
echo 'hadoop' | passwd --stdin hadoop	

master2,3,4节点同上
for i in `seq 2 4`;do ssh root@master${i} "echo 'hadoop' | passwd --stdin hadoop";done

1.4 让master1（主控节点）的hadoop用户能通过密钥登录master1，2,3,4

[root@master1 ~]# su - hadoop

生成密钥和公钥：
[hadoop@master1 ~]$ ssh-keygen -t rsa -P 'hadoop'

将master1的公钥拷贝到master1，2,3,4节点：
[hadoop@master1 ~]$ for i in `seq 1 4`;do ssh-copy-id -i .ssh/id_rsa.pub hadoop@master${i};done

2、hadoop安装

2.1 创建目录并配置权限

[root@master1 ~]# mkdir -pv /bdapps /data/hadoop/hdfs/{nn,snn,dn}

[root@master1 ~]# chown -R hadoop:hadoop /data/hadoop/hdfs/

展开hadoop：
[root@master1 ~]# tar xf hadoop-2.6.2.tar.gz -C /bdapps/

创建软链接：
[root@master1 ~]# cd /bdapps/
[root@master1 bdapps]# ln -sv hadoop-2.6.2 hadoop

创建日志目录，并授权
[root@master1 ~]# cd /bdapps/hadoop
[root@master1 hadoop]# mkdir logs
[root@master1 hadoop]# chmod g+w logs

修改hadoop安装目录权限
[root@master1 hadoop]# chown -R hadoop:hadoop ./*

2.2 主节点（master1）配置

[root@master1 ~]# cd /bdapps/hadoop/etc/hadoop/

[root@master1 hadoop]# vim core-site.xml 

<configuration>
    <property>
        <name>fs.defaultFS</name>
        <value>hdfs://master:8020</value>
        <final>true</final>
    </property>
</configuration>


[root@master1 hadoop]# vim yarn-site.xml 

<configuration>
    <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.address</name>
        <value>master:8032</value>
    </property>
    <property>
        <name>yarn.resourcemanager.scheduler.address</name>
        <value>master:8030</value>
    </property>
        <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>10.201.106.131:8088</value>
    </property>
    <property>
        <name>yarn.nodemanager.aux-services</name>
        <value>mapreduce_shuffle</value>
    </property>
    <property>
        <name>yarn.nodemanager.auxservices.mapreduce_shuffle.class</name>
        <value>org.apache.hadoop.mapred.ShuffleHandler</value>
    </property>
    <property>
        <name>yarn.resourcemanager.scheduler.class</name>
        <value>org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.CapacityScheduler</value>
    </property>
</configuration>


[root@master1 hadoop]# vim hdfs-site.xml 

<configuration>
    <property>
        <name>dfs.replication</name>
        <value>2</value>
    </property>
    <property>
        <name>dfs.namenode.name.dir</name>
        <value>file:///data/hadoop/hdfs/nn</value>
    </property>
    <property>
        <name>dfs.datanode.data.dir</name>
        <value>file:///data/hadoop/hdfs/dn</value>
    </property>
    <property>
        <name>fs.checkpoint.dir</name>
        <value>file:///data/hadoop/hdfs/snn</value>
    </property>
    <property>
        <name>fs.checkpoint.edits.dir</name>
        <value>file:///data/hadoop/hdfs/snn</value>
    </property>
</configuration>


[root@master1 hadoop]# vim mapred-site.xml

<configuration>
    <property>
        <name>mapreduce.framework.name</name>
        <value>yarn</value>
    </property>
</configuration>

[root@master1 hadoop]# vim slaves 

master2
master3
master4

2.2 配置三个从节点

[root@master2 ~]# mkdir -pv /bdapps /data/hadoop/hdfs/dn
[root@master2 ~]# chown -R hadoop:hadoop /data/hadoop/hdfs/

[root@master2 ~]# tar xf hadoop-2.6.2.tar.gz -C /bdapps/

[root@master2 bdapps]# ln -sv hadoop-2.6.2 hadoop

[root@master2 bdapps]# cd hadoop
[root@master2 bdapps]# mkdir logs
[root@master2 bdapps]# chmod g+w logs
[root@master2 bdapps]# chown -R hadoop:hadoop ./*

从master1拷贝配置文件：
[root@master1 hadoop]# su - hadoop

[hadoop@master1 ~]$ scp /bdapps/hadoop/etc/hadoop/* master2:/bdapps/hadoop/etc/hadoop/

2.3 格式化文件系统

[hadoop@master1 ~]$ hdfs namenode -format

2.4 启动mapreduce集群

启动集群datanode节点
[hadoop@master1 ~]$ start-dfs.sh
Starting namenodes on [master]
The authenticity of host 'master (10.201.106.131)' can't be established.
ECDSA key fingerprint is 5e:5d:4d:d2:3f:73:fb:5c:c4:26:c7:c4:85:10:c9:75.
Are you sure you want to continue connecting (yes/no)? yes
master: Warning: Permanently added 'master' (ECDSA) to the list of known hosts.
master: starting namenode, logging to /bdapps/hadoop/logs/hadoop-hadoop-namenode-master1.com.out
master2: starting datanode, logging to /bdapps/hadoop/logs/hadoop-hadoop-datanode-master2.com.out
master4: starting datanode, logging to /bdapps/hadoop/logs/hadoop-hadoop-datanode-master4.com.out
master3: starting datanode, logging to /bdapps/hadoop/logs/hadoop-hadoop-datanode-master3.com.out
Starting secondary namenodes [0.0.0.0]
The authenticity of host '0.0.0.0 (0.0.0.0)' can't be established.
ECDSA key fingerprint is 5e:5d:4d:d2:3f:73:fb:5c:c4:26:c7:c4:85:10:c9:75.
Are you sure you want to continue connecting (yes/no)? yes
0.0.0.0: Warning: Permanently added '0.0.0.0' (ECDSA) to the list of known hosts.
0.0.0.0: starting secondarynamenode, logging to /bdapps/hadoop/logs/hadoop-hadoop-secondarynamenode-master1.com.out

查看各节点启动的进程：
[hadoop@master1 ~]$ jps
4977 NameNode
5324 Jps
5155 SecondaryNameNode

[root@master2 hadoop]# su - hadoop
Last login: Sun Apr 22 11:52:57 CST 2018 from master1 on pts/1
[hadoop@master2 ~]$ 
[hadoop@master2 ~]$ 
[hadoop@master2 ~]$ jps
9972 DataNode
10131 Jps

确认主节点能够连接到另外三个从节点
[root@master1 ~]# netstat -tanp | grep 8020
tcp        0      0 10.201.106.131:8020     0.0.0.0:*               LISTEN      4977/java                            
tcp        0      0 10.201.106.131:8020     10.201.106.134:51956    ESTABLISHED 4977/java           
tcp        0      0 10.201.106.131:8020     10.201.106.133:36426    ESTABLISHED 4977/java           
tcp        0      0 10.201.106.131:8020     10.201.106.132:37988    ESTABLISHED 4977/java 


上传文件测试：
[hadoop@master1 ~]$ hdfs dfs -mkdir /test
[hadoop@master1 ~]$ hdfs dfs -put /etc/fstab /test/fstab
[hadoop@master1 ~]$ hdfs dfs -ls /test/fstab
-rw-r--r--   2 hadoop supergroup       1065 2018-04-23 03:00 /test/fstab
真实文件路径：
[hadoop@master2 logs]$ cat /data/hadoop/hdfs/dn/current/BP-1262978243-10.201.106.131-1524421803827/current/finalized/subdir0/subdir0/blk_1073741827
[hadoop@master4 ~]$ cat /data/hadoop/hdfs/dn/current/BP-1262978243-10.201.106.131-1524421803827/current/finalized/subdir0/subdir0/blk_1073741827

2.5 启动yarn集群

[hadoop@master1 ~]$ start-yarn.sh
#主节点启动了ResourceManager
starting yarn daemons
starting resourcemanager, logging to /bdapps/hadoop/logs/yarn-hadoop-resourcemanager-master1.com.out
master3: starting nodemanager, logging to /bdapps/hadoop/logs/yarn-hadoop-nodemanager-master3.com.out
master4: starting nodemanager, logging to /bdapps/hadoop/logs/yarn-hadoop-nodemanager-master4.com.out
master2: starting nodemanager, logging to /bdapps/hadoop/logs/yarn-hadoop-nodemanager-master2.com.out
[hadoop@master1 ~]$ jps
5919 ResourceManager
4977 NameNode
5155 SecondaryNameNode
6190 Jps

从节点启动了NodeManager
[hadoop@master2 logs]$ jps
10243 DataNode
10508 Jps
10405 NodeManager
[hadoop@master3 ~]$ jps
9380 DataNode
9696 NodeManager
9796 Jps

2.6 查看WEB界面状态

浏览器访问：http://10.201.106.131:8088

浏览器访问：http://10.201.106.131:50070

3、其他操作

3.1 上传大文件观察切块

生成一个200M文件：
[hadoop@master1 ~]$ dd if=/dev/zero of=test bs=1M count=200

原始图：

上传后（超过64M后会切块）：

3.2 通过浏览器查看日志

访问：http://10.201.106.131:50070/logs/

3.2 运行任务测试

列出该测试程序的可用示例：
[hadoop@master1 ~]$ yarn jar /bdapps/hadoop/share/hadoop/mapreduce/hadoop-mapreduce-examples-2.6.2.jar 

统计文件单词数：
[hadoop@master1 ~]$ yarn jar /bdapps/hadoop/share/hadoop/mapreduce/hadoop-mapreduce-examples-2.6.2.jar wordcount /test/fstab /test/functions /test/wordou

查看：

查看任务进度：

查看结果 ：

[hadoop@master1 ~]$ hdfs dfs -ls /test/wordout
Found 2 items
-rw-r--r--   2 hadoop supergroup          0 2018-04-23 06:56 /test/wordout/_SUCCESS
-rw-r--r--   2 hadoop supergroup       7855 2018-04-23 06:56 /test/wordout/part-r-00000

4、yarn集群管理命令

4.1 查看yarn的所有命令

[hadoop@master1 ~]$ yarn

4.2 application

4.2.1 查看作业

列出活动作业：
[hadoop@master1 ~]$ yarn application -list
18/04/23 07:23:47 INFO client.RMProxy: Connecting to ResourceManager at master/10.201.106.131:8032
Total number of applications (application-types: [] and states: [SUBMITTED, ACCEPTED, RUNNING]):0
                Application-Id	    Application-Name	    Application-Type      User	     Queue	             State	       Final-State         Progress	                       Tracking-URL

列出所有作业：
[hadoop@master1 ~]$ yarn application -list -appStates=all
18/04/23 07:24:48 INFO client.RMProxy: Connecting to ResourceManager at master/10.201.106.131:8032
Total number of applications (application-types: [] and states: [NEW, NEW_SAVING, SUBMITTED, ACCEPTED, RUNNING, FINISHED, FAILED, KILLED]):1
                Application-Id	    Application-Name	    Application-Type      User	     Queue	             State	       Final-State         Progress	                       Tracking-URL
application_1524424153008_0001	          word count	           MAPREDUCE    hadoop	   default	          FINISHED	         SUCCEEDED             100%	http://master2:19888/jobhistory/job/job_1524424153008_0001
[hadoop@master1 ~]$ 

查看作业状态：
[hadoop@master1 ~]$ yarn application -status application_1524424153008_0001
18/04/23 07:28:32 INFO client.RMProxy: Connecting to ResourceManager at master/10.201.106.131:8032
Application Report : 
	Application-Id : application_1524424153008_0001
	Application-Name : word count
	Application-Type : MAPREDUCE
	User : hadoop
	Queue : default
	Start-Time : 1524437422005
	Finish-Time : 1524437801216
	Progress : 100%
	State : FINISHED
	Final-State : SUCCEEDED
	Tracking-URL : http://master2:19888/jobhistory/job/job_1524424153008_0001
	RPC Port : 40927
	AM Host : master2
	Aggregate Resource Allocation : 1326835 MB-seconds, 909 vcore-seconds
	Diagnostics : 
[hadoop@master1 ~]$ 

4.3 node

列出node列表：
[hadoop@master1 ~]$ yarn node -list
18/04/23 07:33:37 INFO client.RMProxy: Connecting to ResourceManager at master/10.201.106.131:8032
Total Nodes:3
         Node-Id	     Node-State	Node-Http-Address	Number-of-Running-Containers
   master4:47410	        RUNNING	     master4:8042	                           0
   master3:55126	        RUNNING	     master3:8042	                           0
   master2:54307	        RUNNING	     master2:8042	                           0
列出所有node节点，包括故障下线的：
[hadoop@master1 ~]$ yarn node -list -all


查看指定节点状态信息：
[hadoop@master1 ~]$ yarn node -status master2:54307
18/04/23 07:41:01 INFO client.RMProxy: Connecting to ResourceManager at master/10.201.106.131:8032
Node Report : 
	Node-Id : master2:54307
	Rack : /default-rack
	Node-State : RUNNING
	Node-Http-Address : master2:8042
	Last-Health-Update : Sun 22/Apr/18 10:06:49:900CST
	Health-Report : 
	Containers : 0
	Memory-Used : 0MB
	Memory-Capacity : 8192MB
	CPU-Used : 0 vcores
	CPU-Capacity : 8 vcores
	Node-Labels : 

4.4 logs

在yarn-site.xml配置文件定义yarn.log-aggregation-enable属性的值为true即可。

需要重启集群

查看作业日志
[hadoop@master1 ~]$ yarn logs -applicationId application_1524424153008_0001

4.5 classpath

查看java环境路径：
[hadoop@master1 ~]$ yarn classpath
/bdapps/hadoop/etc/hadoop:/bdapps/hadoop/etc/hadoop:/bdapps/hadoop/etc/hadoop:/bdapps/hadoop/share/hadoop/common/lib/*:/bdapps/hadoop/share/hadoop/common/*:/bdapps/hadoop/share/hadoop/hdfs:/bdapps/hadoop/share/hadoop/hdfs/lib/*:/bdapps/hadoop/share/hadoop/hdfs/*:/bdapps/hadoop/share/hadoop/yarn/lib/*:/bdapps/hadoop/share/hadoop/yarn/*:/bdapps/hadoop/share/hadoop/mapreduce/lib/*:/bdapps/hadoop/share/hadoop/mapreduce/*:/contrib/capacity-scheduler/*.jar:/bdapps/hadoop/share/hadoop/yarn/*:/bdapps/hadoop/share/hadoop/yarn/lib/*

---

yarn的管理命令

4.6 rmadmin

4.6.1 命令帮助

获取命令帮助：
[hadoop@master1 ~]$ yarn rmadmin -help

4.6.2 刷新node节点状态信息

[hadoop@master1 ~]$ yarn rmadmin -refreshNodes
18/04/23 07:54:48 INFO client.RMProxy: Connecting to ResourceManager at master/10.201.106.131:8033

4.7 运行YARN Application流程

1、Application初始化及体检；
2、分配内存并启动AM；
3、AM注册及资源分配；
4、启动并监控容易；
5、Application进度报告；
6、Application运行完成；

5、其他

5.1 官方自动安装部署hadoop工具：Ambari