oozie调度sqoop

oozie调度sqoop

oozie版本4.2.0
sqoop版本1.4.6

调用方式：
restful API方式调用oozie然后触发sqoop shell命令

URL： http://192.168.0.201:11000/oozie/v1/jobs?jobtype=sqoop
请求类型：POST
HTTP content：

<configuration>
    <property>
        <name>fs. default.name</name>
        <value>hdfs: // dn-sa:9000</value>
    </property>
    <property>
        <name>mapred.job.tracker</name>
        <value>dn-sa:8032</value>
    </property>
    <property>
      <name>queueName</name>
      <value> default</value>
    </property>
    <property>
        <name>user.name</name>
        <value>hadoop</value>
    </property>
    <property>
        <name>mapreduce.job.user.name</name>
        <value>hadoop</value>
    </property>
    <property>
        <name>oozie.sqoop.command</name>
        <value>
             import 
            --connect 
            jdbc:mysql: // localhost:3306/oozie 
            --username 
            xxx
            --password 
            xxx
            --table 
            WF_JOBS 
            --target-dir 
            /home/hadoop/oozie-4.2.0-deployed/oozie-jobs/output-data/data_3
        </value>
    </property>
    <property>
    <name>oozie.libpath</name>
    <value>hdfs: // dn-sa:9000/user/hadoop/share/lib/lib_20161010103517/sqoop</value>
    </property>
    <property>
        <name>oozie.proxysubmission</name>
        <value> true</value>
    </property>
</configuration>

说明：
1：上面有个坑爹的地方，上面sqoop命令行之间必须要加换行符，否则oozie提交过去，日志会打出No such sqoop tool: import --connect  xxxxxxxxxxxxxxxx   日志在hadoop的logs文件夹里看。 换句话说代码里面必须加上\r\n换行符
2：另外一个坑爹的地方，必须要把DB的driver添加到$HADOOP_HOME/share/hadoop/yarn目录下， 否则会提示找不到驱动。

至此整个任务调度完成。