Github user koeninger commented on a diff in the pull request:
https://github.com/apache/spark/pull/14026#discussion_r69828832
--- Diff:
external/kafka-0-10/src/main/scala/org/apache/spark/streaming/kafka010/ConsumerStrategy.scala
---
@@ -79,8 +81,71 @@ private case class Subscribe[K, V](
def onStart(currentOffsets: ju.Map[TopicPartition, jl.Long]):
Consumer[K, V] = {
val consumer = new KafkaConsumer[K, V](kafkaParams)
consumer.subscribe(topics)
- if (currentOffsets.isEmpty) {
- offsets.asScala.foreach { case (topicPartition, offset) =>
+ val toSeek = if (currentOffsets.isEmpty) {
+ offsets
+ } else {
+ currentOffsets
--- End diff --
The consumer needs to be in a state that poll() is safe to call on it.
With auto offset reset none, a consumer with no positions will throw if poll is
called. You're right that the driver knows the recovered positions, this is
more about working around consumer state.
---
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.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
