[GitHub] [rocketmq-flink] zhouxinyu commented on a diff in pull request #46: [#35] Use `LitePullConsumer` model instead of default pull consumer

[GitBox]

zhouxinyu commented on code in PR #46:
URL: https://github.com/apache/rocketmq-flink/pull/46#discussion_r992088138


##########
src/main/java/org/apache/rocketmq/flink/legacy/RocketMQSourceFunction.java:
##########
@@ -404,8 +380,7 @@ private long getMessageQueueOffset(MessageQueue mq) throws 
MQClientException {
     private void updateMessageQueueOffset(MessageQueue mq, long offset) throws 
MQClientException {
         offsetTable.put(mq, offset);
         if (!enableCheckpoint) {
-            consumer.updateConsumeOffset(mq, offset);
-            consumer.getOffsetStore().persist(consumer.queueWithNamespace(mq));
+            consumer.getOffsetStore().updateOffset(mq, offset, false);

Review Comment:
   It depends. The consumer will persist the offsets periodically, even if we 
only call `updateOffset`. The `persist` call is heavy, each call will trigger a 
RPC.
   
   So, if the updateMessageQueueOffset already batched some requests, `persist` 
is more suitable. If not and we can tolerate redundant messages in some 
abnormal cases, updateOffset is enough.



-- 
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: dev-unsubscr...@rocketmq.apache.org

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