Github user hmcl commented on a diff in the pull request: https://github.com/apache/storm/pull/2465#discussion_r157831049 --- Diff: external/storm-kafka-client/src/main/java/org/apache/storm/kafka/spout/KafkaSpoutConfig.java --- @@ -115,26 +115,30 @@ public KafkaSpoutConfig(Builder<K, V> builder) { } /** - * The offset used by the Kafka spout in the first poll to Kafka broker. The choice of this parameter will affect the number of consumer - * records returned in the first poll. By default this parameter is set to UNCOMMITTED_EARLIEST. <br/><br/> + * Defines the offset used by the {@link KafkaSpout} in the first poll to Kafka broker. The choice of this parameter will affect + * the number of consumer records returned in the first poll. By default this parameter is set to UNCOMMITTED_EARLIEST. <br/><br/> * The allowed values are EARLIEST, LATEST, UNCOMMITTED_EARLIEST, UNCOMMITTED_LATEST. <br/> * <ul> - * <li>EARLIEST means that the kafka spout polls records starting in the first offset of the partition, regardless of previous - * commits</li> - * <li>LATEST means that the kafka spout polls records with offsets greater than the last offset in the partition, regardless of - * previous commits</li> + * <li>EARLIEST when the topology is first deployed the kafka spout polls records starting in the first offset of the --- End diff -- OK, however this statement is correct because if a topology is deployed again it is a new topology because it has a new id and creates new object instances.
---