RivenSun2 commented on a change in pull request #11340:
URL: https://github.com/apache/kafka/pull/11340#discussion_r713524099



##########
File path: 
clients/src/main/java/org/apache/kafka/clients/consumer/internals/ConsumerCoordinator.java
##########
@@ -1069,6 +1065,37 @@ private void maybeAutoCommitOffsetsSync(Timer timer) {
         }
     }
 
+    private void cleanUpConsumedOffsets(Map<TopicPartition, OffsetAndMetadata> 
willCommitOffsets) {
+
+        if (willCommitOffsets.isEmpty())
+            return;
+
+        Set<String> validTopics = metadata.fetch().topics();
+        Set<TopicPartition> toGiveUpTopicPartitions = new HashSet<>();
+
+        Iterator<Map.Entry<TopicPartition, OffsetAndMetadata>> iterator = 
willCommitOffsets.entrySet().iterator();
+
+        while (iterator.hasNext()) {
+
+            Map.Entry<TopicPartition, OffsetAndMetadata> entry = 
iterator.next();
+
+            if (!validTopics.contains(entry.getKey().topic())) {
+
+                toGiveUpTopicPartitions.add(entry.getKey());
+                iterator.remove();
+            }
+
+        }
+
+        if (toGiveUpTopicPartitions.size() > 0) {
+
+            //Because toGiveUpTopicPartitions may receive 
`UnknownTopicOrPartitionException` when submitting their offsets.
+            //We are prepared to abandon them. The worst effect is that these 
partitions may repeatedly consume some messages
+            log.warn("Synchronous auto-commit of offsets {} will be 
abandoned", toGiveUpTopicPartitions);

Review comment:
       Thank you for your reply.
   Can you simulate the case of jira and debug test it locally?
   
   The `allConsumedOffsets` variable is a collection of all topicPartitions 
with a valid position in the consumer's local area
   sourceCode as below:
   `Map<TopicPartition, OffsetAndMetadata> allConsumedOffsets = 
subscriptions.allConsumed();`
   
   `public synchronized Map<TopicPartition, OffsetAndMetadata> allConsumed() {
           Map<TopicPartition, OffsetAndMetadata> allConsumed = new HashMap<>();
           assignment.forEach((topicPartition, partitionState) -> {
               if (partitionState.hasValidPosition())
                   allConsumed.put(topicPartition, new 
OffsetAndMetadata(partitionState.position.offset,
                           partitionState.position.offsetEpoch, ""));
           });
           return allConsumed;
       }`
   
   
   In fact, what is filtered here are all topicPartitions that are ready to 
commitOffsets but may trigger `UnknownTopicOrPartitionException`




-- 
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-unsubscr...@kafka.apache.org

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


Reply via email to