Github user srdo commented on a diff in the pull request:
https://github.com/apache/storm/pull/2465#discussion_r157349371
--- Diff:
external/storm-kafka-client/src/main/java/org/apache/storm/kafka/spout/KafkaSpout.java
---
@@ -225,6 +237,23 @@ private long doSeek(TopicPartition tp,
OffsetAndMetadata committedOffset) {
}
}
+ /**
+ * Checks If {@link OffsetAndMetadata} was committed by this topology,
either by this or another spout instance.
--- End diff --
Sort of. I believe the storm id will be new every time you deploy a
topology. According to the javadoc for getStormId it's the topology name + a
nonce. So EARLIEST and LATEST will apply when you deploy the topology, but not
when activating/deactivating the topology, and it won't apply if the spout
crashes either.
---
