scwhittle commented on code in PR #32456:
URL: https://github.com/apache/beam/pull/32456#discussion_r1760696533
##########
sdks/java/io/kafka/src/main/java/org/apache/beam/sdk/io/kafka/ReadFromKafkaDoFn.java:
##########
@@ -461,6 +461,11 @@ public ProcessContinuation processElement(
return ProcessContinuation.resume();
}
for (ConsumerRecord<byte[], byte[]> rawRecord : rawRecords) {
+ // If the Kafka consumer returns a record with an offset that is
already processed
Review Comment:
Actually perhaps it is better to just check the position directly after the
seek. That would ensure this logic just covers the bad seek behavior and
doesn't affect other processing (which could mask some other serious bug if
records are arriving out of order).
We can query the consumer position after performign the seek to make sure it
is correct.
https://kafka.apache.org/25/javadoc/org/apache/kafka/clients/consumer/KafkaConsumer.html#position-org.apache.kafka.common.TopicPartition-java.time.Duration-
--
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]
