Github user rmetzger commented on a diff in the pull request:
https://github.com/apache/flink/pull/2108#discussion_r67296678
--- Diff:
flink-streaming-connectors/flink-connector-kafka-base/src/main/java/org/apache/flink/streaming/connectors/kafka/FlinkKafkaProducerBase.java
---
@@ -276,6 +314,41 @@ public void close() throws Exception {
checkErroneous();
}
+ // ------------------- Logic for handling checkpoint flushing
-------------------------- //
+
+ private void acknowledgeMessage() {
+ if(!flushOnCheckpoint) {
+ // the logic is disabled
+ return;
+ }
+ pendingRecords--;
+ }
+
+ @Override
+ public Serializable snapshotState(long checkpointId, long
checkpointTimestamp) {
+ if(flushOnCheckpoint) {
+ // flushing is activated: We need to wait until
pendingRecords is 0
+ while(pendingRecords > 0) {
+ try {
+ Thread.sleep(10);
--- End diff --
The problem is that the `flush()` method is only implemented by the Kafka
0.9 producer, not by the 0.8 implementation.
As you can see from the classname, its the shared base class between the
two version specific implementations. I think for the 0.8 producer, there is no
way around the waiting approach.
I'll update the pull request to call `flush()` on the 0.9 producer.
---
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.
---
