[FLINK-1958] Improve error handling of receiver task failures

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.10.0
• Fix Version/s: 0.9
• Component/s: Runtime / Coordination
• Assignee: Unassigned
• Reporter: Ufuk Celebi
• Created: 29/Apr/15 11:55
• Updated: 06/Oct/15 15:51
• Resolved: 01/Jun/15 07:26

Description

Currently, receiver task failures silently fail.

We need the following behaviour:

• Sender keeps going. May be back-pressured when no receiver pulls the data any more.
• Sender may be cancelled by JobManager
• Partition stays sane
• Netty channel needs to be closed
• Transfer needs to be canceled by a cancel message (receiver to sender)

Issue Links

• is part of: FLINK-1954 Task Failures and Error Handling