[关闭]
@zhangyy 2020-11-19T09:28:06.000000Z 字数 8622 阅读 161

oozie 任务调度处理

协作框架

  • 一：oozie example 运行任务调度案例
  • 二：oozie 运行自定的mapreduce 的jar 包
  • 三：oozie 调度shell 脚本
  • 四：oozie 的coordinator 周期性调度当前任务

一： 运行oozie example 案例

1.1 解压exmaple包

  1. 解压example 包
  2. tar -zxvf oozie-examples.tar.gz
  4. cd /home/hadoop/yangyang/oozie/examples/apps/map-reduce
  6. job.properties      --定义job相关的属性，比如目录路径、namenode节点等。
  7.                     --定义workflow的位置
  9. workflow.xml    --定义工作流相关的配置（start  --end   --kill）（action）
  10.                 --mapred.input.dir
  11.                 --mapred.output.dir
  13. lib     --目录，存放job任务需要的资源（jar包）

1.2 更改job.properties

  1. nameNode=hdfs://namenode01.hadoop.com:8020
  2. jobTracker=namenode01.hadoop.com:8032
  3. queueName=default
  4. examplesRoot=examples
  6. oozie.wf.application.path=${nameNode}/user/hadoop/${examplesRoot}/apps/map-reduce/workflow.xml
  7. outputDir=map-reduce

1.3 配置workflow.xml 文件：

  1. <workflow-app xmlns="uri:oozie:workflow:0.2" name="map-reduce-wf">
  2.     <start to="mr-node"/>
  3.     <action name="mr-node">
  4.         <map-reduce>
  5.             <job-tracker>${jobTracker}</job-tracker>
  6.             <name-node>${nameNode}</name-node>
  7.             <prepare>
  8.                 <delete path="${nameNode}/user/${wf:user()}/${examplesRoot}/output-data/${outputDir}"/>
  9.             </prepare>
  10.             <configuration>
  11.                 <property>
  12.                     <name>mapred.job.queue.name</name>
  13.                     <value>${queueName}</value>
  14.                 </property>
  15.                 <property>
  16.                     <name>mapred.mapper.class</name>
  17.                     <value>org.apache.oozie.example.SampleMapper</value>
  18.                 </property>
  19.                 <property>
  20.                     <name>mapred.reducer.class</name>
  21.                     <value>org.apache.oozie.example.SampleReducer</value>
  22.                 </property>
  23.                 <property>
  24.                     <name>mapred.map.tasks</name>
  25.                     <value>1</value>
  26.                 </property>
  27.                 <property>
  28.                     <name>mapred.input.dir</name>
  29.                     <value>/user/${wf:user()}/${examplesRoot}/input-data/text</value>
  30.                 </property>
  31.                 <property>
  32.                     <name>mapred.output.dir</name>
  33.                     <value>/user/${wf:user()}/${examplesRoot}/output-data/${outputDir}</value>
  34.                 </property>
  35.             </configuration>
  36.         </map-reduce>
  37.         <ok to="end"/>
  38.         <error to="fail"/>
  39.     </action>
  40.     <kill name="fail">
  41.         <message>Map/Reduce failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
  42.     </kill>
  43.     <end name="end"/>
  44. </workflow-app>

1.3 上传example 目录到hdfs 上面

  1. hdfs dfs -put example example

image_1aki0h9sk1qrf1t7l1hjo18b1mll9.png-11.7kB
image_1aki0o964i62lpu1iolnrj1stmm.png-58kB

1.4 运行oozie 调度任务

  1. bin/oozie job -oozie http://namenode01.hadoop.com:11000/oozie -config examples/apps/map-reduce/job.properties -run

image_1aki2bafs1in9s6h1lt11j1m17km.png-15.9kB

查看状态：

image_1aki2apoaup14av14li1ea25959.png-39.1kB

输出目录

image_1aki2i60p1vg41dse1u752aj1u1313.png-66.6kB
image_1aki2j2n71ngsvv81q9o1rtd1k7d1g.png-105.4kB
image_1aki2k7h417e2dfpsj51l1u134a1t.png-109.4kB

二：oozie 运行自定的mapreduce 的jar 包

2.1 在hdfs 上创建上传目录

  1. cd /home/hadoop/yangyang/oozie/
  2. hdfs dfs -mkdir oozie-apps

2.2 新建本地的文件用作上传的目录

  1. mkdir oozie-apps
  2. cd /home/hadoop/yangyang/oozie/examples/apps
  3. cp -ap map-reduce /home/hadoop/yangyang/oozie/oozie-apps/
  5. cd /homme/hadoop/yangyang/oozie/oozie-appps/map-reduce
  6. mkdir input-data

2.3 拷贝运行的jar包与要运行的job 任务的文件

  1. cp -p mr-wordcount.jar yangyang/oozie/oozie-apps/map-reduce/lib/
  2. cp -p /home/hadoop/wc.input ./input-data

2.4 配置job.properties 文件和workflow.xml

vim job.properties

  1. nameNode=hdfs://namenode01.hadoop.com:8020
  2. jobTracker=namenode01.hadoop.com:8032
  3. queueName=default
  4. examplesRoot=oozie-apps/map-reduce
  6. oozie.wf.application.path=${nameNode}/user/hadoop/${examplesRoot}/workflow.xml
  7. outputDir=oozie-reduce

image_1aki8mlep80s134k1q3d1ks21bec2a.png-24kB

vim workflow.xml

  1. <workflow-app xmlns="uri:oozie:workflow:0.2" name="wc-map-reduce">
  2.     <start to="mr-node"/>
  3.     <action name="mr-node">
  4.         <map-reduce>
  5.             <job-tracker>${jobTracker}</job-tracker>
  6.             <name-node>${nameNode}</name-node>
  7.             <prepare>
  8.                 <delete path="${nameNode}/user/hadoop/${examplesRoot}/output-data/${outputDir}"/>
  9.             </prepare>
  10.             <configuration>
  11.                 <property>
  12.                     <name>mapred.job.queue.name</name>
  13.                     <value>${queueName}</value>
  14.                 </property>
  15.                 <!--0 new API-->                
  16.                 <property>
  17.                     <name>mapred.mapper.new-api</name>
  18.                     <value>true</value>
  19.                 </property>
  20.                 <property>
  21.                     <name>mapred.reducer.new-api</name>
  22.                     <value>true</value>
  23.                 </property>
  25.                 <!--1 input-->
  26.                 <property>
  27.                     <name>mapred.input.dir</name>
  28.                     <value>/user/hadoop/${examplesRoot}/input-data</value>
  29.                 </property>         
  31.                 <!--2 mapper class -->              
  32.                 <property>
  33.                     <name>mapreduce.job.map.class</name>
  34.                     <value>org.apache.hadoop.wordcount.WordCountMapReduce$WordCountMapper</value>
  35.                 </property>
  36.                 <property>
  37.                     <name>mapreduce.map.output.key.class</name>
  38.                     <value>org.apache.hadoop.io.Text</value>
  39.                 </property>
  40.                 <property>
  41.                     <name>mapreduce.map.output.value.class</name>
  42.                     <value>org.apache.hadoop.io.IntWritable</value>
  43.                 </property> 
  45.                 <!--3 reduer class -->                  
  46.                 <property>
  47.                     <name>mapreduce.job.reduce.class</name>
  48.                     <value>org.apache.hadoop.wordcount.WordCountMapReduce$WordCountReducer</value>
  49.                 </property>
  50.                 <property>
  51.                     <name>mapreduce.job.output.key.class</name>
  52.                     <value>org.apache.hadoop.io.Text</value>
  53.                 </property>
  54.                 <property>
  55.                     <name>mapreduce.job.output.value.class</name>
  56.                     <value>org.apache.hadoop.io.IntWritable</value>
  57.                 </property>             
  59.                 <!--4 output -->    
  60.                 <property>
  61.                     <name>mapred.output.dir</name>
  62.                     <value>/user/hadoop/${examplesRoot}/output-data/${outputDir}</value>
  63.                 </property> 
  64.             </configuration>
  65.         </map-reduce>
  66.         <ok to="end"/>
  67.         <error to="fail"/>
  68.     </action>
  69.     <kill name="fail">
  70.         <message>Map/Reduce failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
  71.     </kill>
  72.     <end name="end"/>
  73. </workflow-app>

2.6 上传文件到hdfs 上面：

  1. hdfs dfs -put map-reduce oozie-apps

2.7 执行oozie 命令运行job 处理

  1. bin/oozie job -oozie http://namenode01.hadoop.com:11000/oozie -config oozie-apps/map-reduce/job.properties -run

2.8 在浏览器上面查看测试结果

image_1aki9dbt41iko11g92t3ottjn2n.png-16.1kB

image_1aki9e3nh1g581213ss18ru8gk34.png-45.5kB

image_1aki9fcol86j1p4h1pak1r715qp3h.png-50.2kB

image_1aki9gjhj1uprt9j14o7vdd1f9m3u.png-31.3kB

三：oozie 调度shell 脚本

3.1 生成配置文件：

  1. cd /home/hadoop/yangyang/oozie/examples/apps
  2. cp -ap shell/ ../../oozie-apps/
  3. mv shell mem-shell

3.2 书写shell 脚本：

  1. cd /home/hadoop/yangyang/oozie/oozie-apps/mem-shell

vim meminfo.sh

  1. #!/bin/bash
  2. /usr/bin/free -m >> /tmp/meminfo

3.3 配置job.properties 文件和workflow.xml

vim job.properties

  1. nameNode=hdfs://namenode01.hadoop.com:8020
  2. jobTracker=namenode01.hadoop.com:8032
  3. queueName=default
  4. examplesRoot=oozie-apps/mem-shell
  6. oozie.wf.application.path=${nameNode}/user/${user.name}/${examplesRoot}/workflow.xml
  7. EXEC=meminfo.sh

image_1akibh1td13h0fig1jbp1dsfu364b.png-18.2kB

vim workflow.xml

  1. <workflow-app xmlns="uri:oozie:workflow:0.4" name="mem-shell-wf">
  2.     <start to="shell-node"/>
  3.     <action name="shell-node">
  4.         <shell xmlns="uri:oozie:shell-action:0.2">
  5.             <job-tracker>${jobTracker}</job-tracker>
  6.             <name-node>${nameNode}</name-node>
  7.             <configuration>
  8.                 <property>
  9.                     <name>mapred.job.queue.name</name>
  10.                     <value>${queueName}</value>
  11.                 </property>
  12.             </configuration>
  13.             <exec>${EXEC}</exec>
  14.             <file>/user/hadoop/oozie-apps/mem-shell/${EXEC}#${EXEC}</file>
  15.         </shell>
  16.         <ok to="end"/>
  17.         <error to="fail"/>
  18.     </action>
  19.     <kill name="fail">
  20.         <message>Shell action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
  21.     </kill>
  22.     <end name="end"/>
  23. </workflow-app>

3.4 上传配置文件到hdfs 上面

  1. cd /home/hadoop/yangyang/oozie/oozie-apps
  2. hdfs dfs  -put mem-shell oozie-apps

image_1akibnkmg1svtvt9ktu4m2u6k4o.png-28.4kB

3.5 执行oozie 调度 shell脚本

  1. bin/oozie job -oozie http://namenode01.hadoop.com:11000/oozie -config oozie-apps/mem-shell/job.properties -run

image_1akibqg8i1j0o13ei15321d8v1f4h55.png-31.2kB

image_1akibr84i1sfm10qtq7ler9183b5i.png-64.5kB

image_1akibru3n1i2e1458jqb1f0611335v.png-109.8kB

image_1akibsedf1c6qm19clapmnsb6c.png-83.1kB

四：oozie 的coordinator 周期性调度当前任务

4.1 配置时区 更改oozie 的配置文件

  1. cd /home/hadoop/yangyang/oozie/conf
  3. vim oozie-site.xml 增加：
  5.     <property>
  6.         <name>oozie.processing.timezone</name>
  7.         <value>GMT+0800</value>
  8.     </property>
  9.         <property>
  10.         <name>oozie.service.coord.check.maximum.frequency</name>
  11.         <value>false</value>
  12.     </property>

4.2 更改本地 时间

  1.  使用root 账户 配置
  3.  cp -p /etc/localtime /etc/localtime.bak 
  5.  rm -rf /etc/localtime
  7.  cd /usr/share/zoneinfo/Asia/
  9.  cp -p Shanghai /etc/localtime

4.3 更改oozie-consle.js 文件

  1. cd /home/hadoop/yangyang/oozie/oozie-server/webapps/oozie
  3. vim oozie-console.js 
  5. function getTimeZone() {
  6.     Ext.state.Manager.setProvider(new Ext.state.CookieProvider());
  7.     return Ext.state.Manager.get("TimezoneId","GMT+0800");
  8. }

image_1akie7ik4cna40p1jlvmnd65n6p.png-8.1kB

4.4 从新启动oozie 服务

  1. bin/oozie-stop.sh 
  2. bin/oozie-start.sh

4.5 查看oozie 的当前时间

image_1akiebr9214d01ihe1oi1uvr9gp76.png-46.5kB

4.6 配置job.properties 文件和workflow.xml

  1. cd /home/hadoop/yangyang/oozie/examples/apps
  3. cp -ap cron ../../oozie-apps/
  5. cd cron
  7. rm -rf job.properties workflow.xml
  8. cd /home/hadoop/yangyang/oozie/oozie-apps/mem-shell
  10. cp -p * ../cron

配置job.properties

  1. vim job.properties
  3. ---
  4. nameNode=hdfs://namenode01.hadoop.com:8020
  5. jobTracker=namenode01.hadoop.com:8032
  6. queueName=default
  7. examplesRoot=oozie-apps/cron
  9. oozie.coord.application.path=${nameNode}/user/hadoop/${examplesRoot}/
  10. start=2016-06-6T16:57+0800
  11. end=2016-06-6T20:00+0800
  12. workflowAppUri=${nameNode}/user/hadoop/${examplesRoot}/
  13. EXEC=meminfo.sh

image_1akihapdt1m3f5a13n71q2vg4g7j.png-18.8kB

配置workflow.xml

  1. vim workflow.xml
  3. ---
  5. <workflow-app xmlns="uri:oozie:workflow:0.4" name="memcron-shell-wf">
  6.     <start to="shell-node"/>
  7.     <action name="shell-node">
  8.         <shell xmlns="uri:oozie:shell-action:0.2">
  9.             <job-tracker>${jobTracker}</job-tracker>
  10.             <name-node>${nameNode}</name-node>
  11.             <configuration>
  12.                 <property>
  13.                     <name>mapred.job.queue.name</name>
  14.                     <value>${queueName}</value>
  15.                 </property>
  16.             </configuration>
  17.             <exec>${EXEC}</exec>
  18.             <file>/user/hadoop/oozie-apps/cron/${EXEC}#${EXEC}</file>
  19.         </shell>
  20.         <ok to="end"/>
  21.         <error to="fail"/>
  22.     </action>
  23.     <kill name="fail">
  24.         <message>Shell action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
  25.     </kill>
  26.     <end name="end"/>
  27. </workflow-app>

配置coordinator.xml

  1. vim coordinator.xml
  3. ---
  5. <coordinator-app name="cron-coord" frequency="${coord:minutes(2)}" start="${start}" end="${end}" timezone="GMT+0800"
  6.                  xmlns="uri:oozie:coordinator:0.2">
  7.         <action>
  8.         <workflow>
  9.             <app-path>${workflowAppUri}</app-path>
  10.             <configuration>
  11.                 <property>
  12.                     <name>jobTracker</name>
  13.                     <value>${jobTracker}</value>
  14.                 </property>
  15.                 <property>
  16.                     <name>nameNode</name>
  17.                     <value>${nameNode}</value>
  18.                 </property>
  19.                 <property>
  20.                     <name>queueName</name>
  21.                     <value>${queueName}</value>
  22.                 </property>
  23.                                 <property>
  24.                     <name>EXEC</name>
  25.                     <value>${EXEC}</value>
  26.                 </property>
  27.             </configuration>
  28.         </workflow>
  29.     </action>
  30. </coordinator-app>

4.7 上传配置文件到hdfs 上面：

  1. hdfs dfs -put cron oozie-apps

4.8 执行 oozie 命令 运行job

  1. bin/oozie job -oozie http://namenode01.hadoop.com:11000/oozie -config oozie-apps/cron/job.properties -run

4.9 从web浏览job的相关问题

image_1akihodpprn31lo59ol1h1d1qk380.png-38.8kB

image_1akihpt0opo1rp81ujv1p6a5h48d.png-76.7kB

添加新批注
在作者公开此批注前，只有你和作者可见。
回复批注