[SPARK-17370] Shuffle service files not invalidated when a slave is lost

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.0.1, 2.1.0
• Component/s: Spark Core
• Assignee: Eric Liang
• Reporter: Eric Liang
• Created: 02/Sep/16 01:21
• Updated: 22/May/17 03:19
• Resolved: 07/Sep/16 19:34

Description

DAGScheduler invalidates shuffle files when an executor loss event occurs, but not when the external shuffle service is enabled. This is because when shuffle service is on, the shuffle file lifetime can exceed the executor lifetime.

However, it doesn't invalidate shuffle files when the shuffle service itself is lost (due to whole slave loss). This can cause long hangs when slaves are lost since the file loss is not detected until a subsequent stage attempts to read the shuffle files.

Issue Links

• is related to: SPARK-20832 Standalone master should explicitly inform drivers of worker deaths and invalidate external shuffle service outputs
• relates to: SPARK-17519 [MESOS] Enhance robustness when ExternalShuffleService is broken
• links to: [Github] Pull Request #14931 (ericl)