[FLINK-12364] Introduce a CheckpointFailureManager to centralized manage checkpoint failure

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.9.0
• Component/s: Runtime / Checkpointing
• Labels: pull-request-available

Release Note

The method `CheckpointIDCounter#get` has been added to the `CheckpointIDCounter` interface. The method retrieves the current checkpoint ID. This is an API breaking change and will require you to update your own `CheckpointIDCounter` implementation if you are using a custom `HighAvailabilityServices` implementation.

• Assignee: vinoyang
• Reporter: vinoyang
• Created: 29/Apr/19 09:39
• Updated: 23/Aug/21 09:39
• Resolved: 12/Aug/19 09:44

Description

This issue tracks the work of T2 section about in design document : https://docs.google.com/document/d/1ce7RtecuTxcVUJlnU44hzcO2Dwq9g4Oyd8_biy94hJc/edit?usp=sharing

Issue Links

• blocks: FLINK-12514 Refactor the failure checkpoint counting mechanism with ordered checkpoint id
• causes: FLINK-20675 Asynchronous checkpoint failure would not fail the job anymore
• causes: FLINK-17351 CheckpointCoordinator and CheckpointFailureManager ignores checkpoint timeouts
• duplicates: FLINK-4810 Checkpoint Coordinator should fail ExecutionGraph after "n" unsuccessful checkpoints
• relates to: FLINK-23916 Add documentation for how to configure the CheckpointFailureManager
• links to: GitHub Pull Request #8322