[HADOOP-5883] TaskMemoryMonitorThread might shoot down tasks even if their processes momentarily exceed the requested memory

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.20.1
• Hadoop Flags: Reviewed
• Assignee: Hemanth Yamijala
• Reporter: Hemanth Yamijala
• Created: 21/May/09 12:53
• Updated: 08/Jul/09 16:53
• Resolved: 28/May/09 06:38

Description

Currently the TaskMemoryMonitorThread kills tasks as soon as it detects they are consuming more memory than the max value specified. There are valid cases (see HADOOP-5059) where if a program is executed from the task, it might momentarily occupy twice the amount of memory for a short time. Ideally the monitoring thread should handle this case.

Issue Links

• is blocked by: HADOOP-5881 Simplify configuration related to task-memory-monitoring and memory-based scheduling