[HIVE-19480] Implement and Incorporate MAPREDUCE-207

• Type: New Feature
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 1.2.3
• Component/s: HiveServer2
• Assignee: Unassigned
• Reporter: David Mollitor
• Created: 09/May/18 20:55
• Updated: 09/May/18 21:28

Description

• HiveServer2 has the ability to run many MapReduce jobs in parallel.
• Each MapReduce application calculates the job's file splits at the client level
• = HiveServer2 loading many file splits at the same time, putting pressure on memory

"The client running the job calculates the splits for the job by calling getSplits(), then sends them to the application master, which uses their storage locations to schedule map tasks that will process them on the cluster."

• "Hadoop: The Definitive Guide"

MAPREDUCE-207 should address this memory pressure by moving split calculations into ApplicationMaster. Spark and Tez already take this approach.

Once MAPREDUCE-207 is completed, leverage the capability in HiveServer2.

Issue Links

• depends upon: MAPREDUCE-207 Computing Input Splits on the MR Cluster