chia7712 commented on code in PR #19489:
URL: https://github.com/apache/kafka/pull/19489#discussion_r2644748236
##########
clients/src/main/java/org/apache/kafka/clients/producer/internals/Sender.java:
##########
@@ -174,43 +174,6 @@ private void maybeRemoveAndDeallocateBatch(ProducerBatch
batch) {
this.accumulator.deallocate(batch);
}
- /**
- * Get the in-flight batches that has reached delivery timeout.
- */
- private List<ProducerBatch> getExpiredInflightBatches(long now) {
Review Comment:
I understand that network uncertainty is unavoidable, as you mentioned.
However, there is a key distinction between unavoidable network ambiguity and
intentional state divergence. We are actively creating a scenario where the
client logic 'gives up' while the (kafka) I/O layer continues to process the
request
That said, it appears intentional divergence was indeed anticipated by the
KIP-91.
> This implies that user might be notified of batch expiry while a batch is
still in-flight.
While I fail to see the practical benefit of this use case, I am amenable to
preserving this behavior since it has been documented and released for a long
time.
--
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]
