[FLINK-26359] Aligning the retry mechanism for the repeatable cleanup and task restarts

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 1.15.0
• Component/s: API / Core, Runtime / Configuration, Runtime / Coordination
• Assignee: Unassigned
• Reporter: Matthias Pohl
• Created: 24/Feb/22 18:32
• Updated: 24/Feb/22 18:33

Description

Flink has two different classes for modelling retry strategies: org.apache.flink.util.concurrent.RetryStrategy and org.apache.flink.runtime.executiongraph.failover.flip1.RestartBackoffTimeStrategy. Both should be aligned. This would enable us to use the same retry strategies in different parts of the code and having an aligned user experience for this kind of logic.

See org.apache.flink.configuration.CleanupOptions and org.apache.flink.configuration.RestartStrategyOptions! Both provide redundant functionality to provide the same user experience. This should be cleaned up.

Issue Links

• is caused by: FLINK-26331 Make max retries configurable