shuwenwei opened a new pull request, #17941: URL: https://github.com/apache/iotdb/pull/17941
## Description This PR fixes fast compaction writer checkpointing for variable-length data. The previous implementation checked chunk size primarily by point-count checkpoints. For large TEXT/STRING/BLOB/OBJECT values, the writer could accumulate a large PageWriter/ChunkWriter buffer before the next point checkpoint. This patch adds per-subtask written-point-size checkpoint tracking and lets size checkpoints trigger the existing chunk-size check. It also keeps checkpoint state isolated per sub compaction task and updates size accounting for direct page-flush paths so large binary pages can trigger checks promptly. ## Changes - Track written point total size per subtask in compaction writers. - Trigger chunk-size checks when either the point checkpoint or size checkpoint is reached. - Enable size checkpointing only for chunk writers containing binary types. - Account for direct non-aligned and aligned page flushes using serialized page sizes. - Add a focused unit test proving binary-size checkpoints trigger chunk-size checks before point checkpoints. ## Validation - Ran `git diff --cached --check` before commit. - Ran `git diff --check` during local verification. - Did not run Maven, spotless, or formatting checks. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
