RockteMQ-AI commented on issue #10615: URL: https://github.com/apache/rocketmq/issues/10615#issuecomment-4958823149
**Issue Evaluation** Category: `type/bug` | Status: **Confirmed** The reported issue has been verified against the current codebase (develop branch). **Root Cause:** `RocksDBConsumeQueueStore` constructor unconditionally sets `notifyMessageArriveInBatch=true` on the shared `DefaultMessageStore` (`RocksDBConsumeQueueStore.java:105-107`). When `CombineConsumeQueueStore` instantiates an inner `RocksDBConsumeQueueStore` (`CombineConsumeQueueStore.java:97-101`), this globally disables reput-time notification for **all** topics — including file-based CQ topics that are not double-written to RocksDB. **Impact:** Under selective double-write (`rocksdbCQSelectiveDoubleWriteEnable=true`), non-LITE topics fall into a notification dead zone: - `doReput` skips notification (flag is globally `true`) — `DefaultMessageStore.java:2749-2751` - File-based `ConsumeQueueStore` does not call `notifyMessageArriveIfNecessary` - `RocksGroupCommitService` never sees these topics (selective skip at `CombineConsumeQueueStore.java:363-371`) - Long-polling consumers on non-double-written topics only wake up on polling timeout **Severity:** Medium — increases consume latency for affected topics under CombineConsumeQueue + selective double-write configuration. An automated fix proposal will be generated. Reply `/approve` to proceed with PR generation. --- *Automated evaluation by RockteMQ-AI* -- 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]
