Thanks Robert.I think I crossed past the issue by putting
oozie.launcher.mapreduce.map.memory.mb property in the workflow.

On 9/5/14, 5:06 PM, "Robert Kanter" <[email protected]> wrote:

>Hi Sathi,
>
>Is it the launcher job (i.e. your driver code) that's getting the OOM or
>the MR job itself?
>In general, the properties in <configuration> get applied only to the MR
>job.  If you want them to be applied to the launcher job, you should
>prefix
>the properties with "oozie.launcher" (e.g.
>"oozie.launcher.mapreduce.map.memory.mb").
>
>
>Also, with the Java action, if you want the <configuration> properties
>from
>the workflow XML to get passed to your driver for launching the MR job,
>you
>have to load an XML file prepared by Oozie and use that for your MR job's
>JobConf.  It's not currently documented, but you can look at OOZIE-1969
><https://issues.apache.org/jira/browse/OOZIE-1696> for details.
>
>- Robert
>
>
>
>On Fri, Sep 5, 2014 at 4:41 PM, Sathi Chowdhury
><[email protected]
>> wrote:
>
>>
>> >Trying to run a java action (this is the driver class for running a map
>> >reduce job)
>>
>> Here is my workflow xml
>> <action name ="loaddata">
>>         <java>
>>         <job-tracker>${jobTracker}</job-tracker>
>>         <name-node>${nameNode}</name-node>
>>         <job-xml>${workflowRoot}/hive-site.xml</job-xml>
>>         <configuration>
>>         <property>
>>                   <name>mapred.job.queue.name</name>
>>                   <value>${queueName}</value>
>>                 </property>
>>                  <property>
>>                    <name>oozie.action.max.output.data</name>
>>                    <value>100000</value>
>>                </property>
>>   <property>
>>     <name>mapreduce.reduce.memory.mb</name>
>>     <value>4096</value>
>>   </property>
>>   <property>
>>     <name>mapreduce.reduce.java.opts</name>
>>     <value>-Xms4g -Xmx4g</value>
>>   </property>
>>   <property>
>>     <name>mapreduce.map.memory.mb</name>
>>     <value>4096</value>
>>   </property>
>>   <property>
>>     <name>mapreduce.map.java.opts</name>
>>     <value>-Xmx4g -Xms4g</value>
>>   </property>
>> <property>
>>   <name>oozie.launcher.mapred.child.java.opts</name>
>> <value>-Xmx2048m</value>
>> </property>
>> <property>
>>     <name>hcat.metastore.maxpartitions</name>
>>     <value>200000</value>
>>   </property>
>>         </configuration>
>>         <main-class>x.y.z.ParseXTables</main-class>
>>        <java-opts>-Xmx4g -Xms4g</java-opts>
>>         <arg>-conf</arg>
>>         <arg>defaultconf.xml</arg>
>> <file>${workflowRoot}/hive-site.xml</file>
>>         <file>lib/${EXEC}#${EXEC}</file>
>>         <file>${workflowRoot}/defaultconf.xml</file>
>>         <capture-output/>
>>         </java>
>>          <ok to="mark-daily-job-success" />
>>          <error to="fail" />
>>         </action>
>>
>>
>>
>> I am getting this error in the attempt log, I am thinking this happens
>> when oozie launcher tries to run
>>
>> From the job history the job_1409808477902_0089_conf.xml  shows
>> mapreduce.map.memory.mb 2151
>>
>>
>> Wondering why did it not get propagated from the oozie workflow setting
>> and should have been 4096.
>>
>> Not sure what else I have missied.
>> Will appreciate any clue on this.
>> Thanks
>> Sathi
>>
>>
>>
>>

Reply via email to