[MAPREDUCE-2605] Hadoop Streaming (StreamJob) does not delete temporary job/package jar

• Type: Bug
• Status: Resolved
• Priority: Minor
• Resolution: Not A Problem
• Affects Version/s: 0.20.2
• Fix Version/s: 0.22.0
• Assignee: Unassigned
• Reporter: Greg Wittel
• Created: 06/May/11 16:31
• Updated: 18/Jun/11 17:18
• Resolved: 18/Jun/11 17:18

Description

The streaming job driver (org.apache.hadoop.streaming.StreamJob) does not delete the temporary JAR file it generates after a job completes. Without the fix, /var/tmp fills up with streaming job jars until they get wiped.

The jar name is stored in the class variable 'jar_'. The JAR is generated in 'packageJobJar()' and the name stored in jar_.

Fix: run()/submitAndMonitorJob() should clean up the jar_ file when done. Or the JAR could be generatd as a tempfile and cleaned up automatically.