[
https://issues.apache.org/jira/browse/KAFKA-15946?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Kirk True updated KAFKA-15946:
------------------------------
Parent: (was: KAFKA-14246)
Issue Type: Improvement (was: Sub-task)
> AsyncKafkaConsumer should retry commits on the application thread instead of
> auto-retry
> ---------------------------------------------------------------------------------------
>
> Key: KAFKA-15946
> URL: https://issues.apache.org/jira/browse/KAFKA-15946
> Project: Kafka
> Issue Type: Improvement
> Components: clients, consumer
> Reporter: Philip Nee
> Assignee: Kirk True
> Priority: Major
> Labels: consumer-threading-refactor, kip-848-client-support
> Fix For: 3.8.0
>
>
> The original design was that the network thread always completes the future
> whether succeeds or fails. However, in the current patch, I mis-added
> auto-retry functionality because commitSync wasn't retrying. What we should
> be doing is, the commit sync API should catch the RetriableExceptions and
> resend another commit until timesout.
>
> {code:java}
> if (error.exception() instanceof RetriableException) {
> log.warn("OffsetCommit failed on partition {} at offset {}: {}", tp, offset,
> error.message());
> handleRetriableError(error, response);
> retry(responseTime); <--- We probably shouldn't do this.
> return;
> } {code}
>
> {code:java}
> @Override
> public void commitSync(Map<TopicPartition, OffsetAndMetadata> offsets,
> Duration timeout) {
> acquireAndEnsureOpen();
> long commitStart = time.nanoseconds();
> try
> { CompletableFuture<Void> commitFuture = commit(offsets, true); <-- we
> probably should retry here ConsumerUtils.getResult(commitFuture,
> time.timer(timeout)); }
> finally
> { wakeupTrigger.clearTask();
> kafkaConsumerMetrics.recordCommitSync(time.nanoseconds() - commitStart);
> release(); }
> } {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
