[FLINK-4976] Add a way to abort in flight checkpoints

• Type: New Feature
• Status: Closed
• Priority: Major
• Resolution: Implemented
• Affects Version/s: 1.1.3
• Fix Version/s: 1.1.4, 1.2.0
• Component/s: Runtime / State Backends
• Assignee: Stephan Ewen
• Reporter: Stephan Ewen
• Created: 31/Oct/16 16:49
• Updated: 21/Nov/18 16:51
• Resolved: 08/Nov/16 23:13

Description

Currently, a started checkpoint alignment can only be dissolved if a newer checkpoint comes, or an input becomes closed.

To cleanly signal downstream tasks that a checkpoint will not be completed, we need to introduce CancellationBarriers that can be injected into the stream and dissolve a pending stream alignment and notify the task that a pending checkpoint was skipped.

Issue Links

• blocks: FLINK-4975 Add a limit for how much data may be buffered during checkpoint alignment
• relates to: FLINK-10966 Optimize the release blocking logic in BarrierBuffer

Sub-Tasks

• Add Cancellation Barriers to BarrierTracker and BarrierBuffer: Closed
• Report Declined/Canceled Checkpoints to Checkpoint Coordinator: Closed