getting _SUCCESS file for hive script [duplicate]
Question
Possible Duplicate:
Override hadoop’s mapreduce.fileoutputcommitter.marksuccessfuljobs in oozie
i am trying to run a hive script from an oozie workflow. the "mapreduce.fileoutputcommitter.marksuccessfuljobs" property has been set to true but he hive script overwrites it and hence, i dont get the success file. is there a way i can configure the same property in my hive script or any other way to overcome this situation?
<property>
<name>mapreduce.fileoutputcommitter.marksuccessfuljobs</name>
<value>true</value>
</property>
in the job tracker, it says that the "mapreduce.fileoutputcommitter.marksuccessfuljobs" property for hive is set to true but still no _success file is available in the final output location
Solution
This isn't possible because hive overrides the OutputCommitter with it's own NullOutputCommitter. See dupe question here:
Override hadoop's mapreduce.fileoutputcommitter.marksuccessfuljobs in oozie