Github user hmcl commented on a diff in the pull request:

    https://github.com/apache/storm/pull/1924#discussion_r106557594
  
    --- Diff: 
external/storm-kafka-client/src/main/java/org/apache/storm/kafka/spout/KafkaSpout.java
 ---
    @@ -270,17 +272,40 @@ public void setWaitingToEmit(ConsumerRecords<K,V> 
consumerRecords) {
     
         // ======== poll =========
         private ConsumerRecords<K, V> pollKafkaBroker() {
    -        doSeekRetriableTopicPartitions();
    -        if (refreshSubscriptionTimer.isExpiredResetOnTrue()) {
    -            kafkaSpoutConfig.getSubscription().refreshAssignment();
    +        Set<TopicPartition> retriableTopicPartitions = 
doSeekRetriableTopicPartitions();
    +        Set<TopicPartition> partitionsToPause = 
getPartitionsToPauseToEnforceUncommittedOffsetsLimit(retriableTopicPartitions);
    +        try{
    +            if(!partitionsToPause.isEmpty()) {
    +                kafkaConsumer.pause(partitionsToPause);
    +            }
    +            if (refreshSubscriptionTimer.isExpiredResetOnTrue()) {
    +                kafkaSpoutConfig.getSubscription().refreshAssignment();
    +            }
    +            final ConsumerRecords<K, V> consumerRecords = 
kafkaConsumer.poll(kafkaSpoutConfig.getPollTimeoutMs());
    +            final int numPolledRecords = consumerRecords.count();
    +            LOG.debug("Polled [{}] records from Kafka. [{}] uncommitted 
offsets across all topic partitions", numPolledRecords, numUncommittedOffsets);
    +            return consumerRecords;
    +        } finally {
    +            if (!partitionsToPause.isEmpty()) {
    +                //Ensure that we resume all partitions in case some were 
paused when processing retries
    +                kafkaConsumer.resume(kafkaConsumer.assignment());
    +            }
             }
    -        final ConsumerRecords<K, V> consumerRecords = 
kafkaConsumer.poll(kafkaSpoutConfig.getPollTimeoutMs());
    -        final int numPolledRecords = consumerRecords.count();
    -        LOG.debug("Polled [{}] records from Kafka. [{}] uncommitted 
offsets across all topic partitions", numPolledRecords, numUncommittedOffsets);
    -        return consumerRecords;
         }
    -
    -    private void doSeekRetriableTopicPartitions() {
    +    
    +    private Set<TopicPartition> 
getPartitionsToPauseToEnforceUncommittedOffsetsLimit(Set<TopicPartition> 
retriableTopicPartitions) {
    --- End diff --
    
    To make the code more readable, let's put a method comment explaining the 
idea and replace this method with something along the lines:
    ```
    private void 
pauseNonRetriableTpsToEnforceUncommitedOffsetsLimit(Set<TopicPartition> 
retriableTopicPartitions) {
            final Set<TopicPartition> partitionsToPause = new 
HashSet<>(kafkaConsumer.assignment());
            partitionsToPause.removeAll(retriableTopicPartitions);
            kafkaConsumer.pause(partitionsToPause);
            // no need to return anything
        }
    ```


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---

Reply via email to