[KAFKA-18166] Indefinite timeout when publishing events to RLMM (RemoteLogMetadataManager)

• Type: Improvement
• Status: Open
• Priority: Minor
• Resolution: Unresolved
• Affects Version/s: 4.0.0, 3.6.0, 3.7.0, 3.8.0, 3.9.0
• Component/s: Tiered-Storage
• Labels: newbie
• Assignee: PoAn Yang
• Reporter: Divij Vaidya
• Created: 05/Dec/24 15:06
• Updated: 4 days ago 17:25

Description

RemoteLogManager has multiple instances where we are doing a CompletableFuture.get() without setting its timeout parameter. Ideally, we do not want indefinite wait in the code base which depends on external components such as RemoteLogMetadataManager. For example, the code would "hang" if the underlying RLMM is not available or very slow for some reason.

The objective of this task is to set a reasonable timeout (my suggestion is 1 sec) on these futures. You will also need to handle the ensuing TimeoutException accordingly.

https://github.com/apache/kafka/blob/50b6953661a46d7d57a8aca5c875e91a19166253/core/src/main/java/kafka/log/remote/RemoteLogManager.java#L572 

Issue Links

• links to: GitHub Pull Request #18097