echooymxq opened a new issue, #10432:
URL: https://github.com/apache/rocketmq/issues/10432

   ### Before Creating the Enhancement Request
   
   - [x] I have confirmed that this should be classified as an enhancement 
rather than a bug/feature.
   
   
   ### Summary
   
   Currently, when RocketMQ uses RocksDB-based ConsumeQueue, the default read 
path is controlled by `iteratorWhenUseRocksdbConsumeQueue`, which is enabled by 
default.
   
   However, this iterator is a RocketMQ-level reusable iterator. Internally, it 
still calls `RocksDBConsumeQueueTable#rangeQuery`, which builds N RocksDB keys 
for consecutive ConsumeQueue offsets and then calls `multiGet`.
   
   This enhancement proposes optimizing the RocksDB ConsumeQueue sequential 
read path by using a bounded `RocksIterator` scan for consecutive offsets, 
while preserving the existing external ConsumeQueue iterator semantics.
   
   ### Motivation
   
   Using RocksDB's native iterator scan for continuous offsets to improve 
performance.
   
   ### Describe the Solution You'd Like
   
   Use a bounded RocksIterator scan in RocksDBConsumeQueueTable#rangeQuery to 
read continuous offsets while keeping existing ConsumeQueue iterator semantics 
unchanged.
   
   ### Describe Alternatives You've Considered
   
   No
   
   ### Additional Context
   
   _No response_


-- 
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]

Reply via email to