Kafka spark directStream can not get data

问题

I'm using spark directStream api to read data from Kafka. My code as following please:

val sparkConf = new SparkConf().setAppName("testdirectStreaming")
val sc = new SparkContext(sparkConf)
val ssc = new StreamingContext(sc, Seconds(2))

val kafkaParams = Map[String, String](
    "auto.offset.reset" -> "smallest",
    "metadata.broker.list"->"10.0.0.11:9092",
    "spark.streaming.kafka.maxRatePerPartition"->"100"
)
//I set all of the 3 partitions fromOffset are 0
var fromOffsets:Map[TopicAndPartition, Long] = Map(TopicAndPartition("mytopic",0) -> 0)
fromOffsets+=(TopicAndPartition("mytopic",1) -> 0)
fromOffsets+=(TopicAndPartition("mytopic",2) -> 0)

val kafkaData = KafkaUtils.createDirectStream[String, String, StringDecoder, StringDecoder, MessageAndMetadata[String, String]](
ssc, kafkaParams, fromOffsets,(mmd: MessageAndMetadata[String, String]) => mmd)

var offsetRanges = Array[OffsetRange]()
kafkaData.transform { rdd =>
    offsetRanges = rdd.asInstanceOf[HasOffsetRanges].offsetRanges
    rdd
}.map {
    _.message()
}.foreachRDD { rdd =>
    for (o <- offsetRanges) {
        println(s"---${o.topic} ${o.partition} ${o.fromOffset} ${o.untilOffset}")
    }
    rdd.foreachPartition{ partitionOfRecords =>
        partitionOfRecords.foreach { line =>
            println("===============value:"+line)
        }
    }
}

I'm sure there are data in the kafka cluster, but my code could not get any of them. Thanks in advance.

回答1:

I found the reason: The old messages in kafka have already been deleted since the retention period expired. So when I set the fromOffset is 0 it caused OutOfOffSet exception. The exception caused Spark reset the offset with the latest ones. Therefore I could not get any messages. The solution is that I need to set the appropriate fromOffset to avoid the Exception.

来源：https://stackoverflow.com/questions/31978998/kafka-spark-directstream-can-not-get-data

标签

apache-spark

apache-kafka

spark-streaming