[PR] KAFKA-21317: Respect log start offset in max timestamp lookups [kafka]

["shantanukawale-stripe (via GitHub)" <gi...@apache.org>]

shantanukawale-stripe opened a new pull request, #24027:
URL: https://github.com/apache/kafka/pull/24027

   After the log start offset advances, `MAX_TIMESTAMP` can return a record below that offset because segment maxima still include logically deleted records.
   
   Exclude segments before the start offset. Preserve the cached lookup when the winning segment is fully retained; otherwise compare eligible batch maxima and scan retained records in a batch spanning the start offset. Keep the decompressed-record size limit and return empty when no timestamped records remain.
   
   Follow-up to #23880 under [KAFKA-21317](https://issues.apache.org/jira/browse/KAFKA-21317). This addresses the lower offset bound; #23780 separately addresses the fetchable upper bound.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: jira-unsubscribe@kafka.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

Re: [PR] KAFKA-21317: Respect log start offset in max timestamp lookups [kafka]

["shantanukawale-stripe (via GitHub)" <gi...@apache.org>]

shantanukawale-stripe commented on PR #24027:
URL: https://github.com/apache/kafka/pull/24027#issuecomment-6521388450

   Validation on Java 17:
   
   - Full `UnifiedLogTest`: 215 tests passed.
   - `:storage:checkstyleMain`, `:storage:checkstyleTest`, `:storage:spotlessCheck`, and `:storage:spotbugsMain` passed.
   - The seven initial regression cases fail against unmodified trunk and pass with this change. Additional coverage checks that expired oversized batches are skipped and retained oversized records still enforce the decompression limit.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: jira-unsubscribe@kafka.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org