kafka-spark偏移量提交至redis kafka1.0版本

kafka版本 1.0.0

spark版本 spark-streaming-kafka-0-10_2.11
/**
* @created by imp ON 2019/12/21
*/
class KafkaManagerByRedis(zkHost:String,kafkaParams: Map[String, Object]) extends Logging {

private val (zkClient,zkConnection) = ZkUtils.createZkClientAndConnection(zkHost , 10000 , 10000)
private val zkUtils = new ZkUtils(zkClient,zkConnection , false)
private val jedis = JedisUtil.getInstance().getJedis

/**
* def createDirectStream:InputDStream
**/

def createDirectStream[K: ClassTag, V: ClassTag](ssc: StreamingContext, topics: Seq[String]): InputDStream[ConsumerRecord[K, V]] = {
//1:readOffset
val groupId = kafkaParams("group.id").toString
val topic = topics(0)
val topicPartition: Map[TopicPartition, Long] = readOffset(topic, groupId)
KafkaUtils.createDirectStream[K, V](
ssc,
PreferConsistent,
Subscribe[K, V](topics, kafkaParams, topicPartition)
)
}

/**
* 读取偏移量
*
* @param topics
* @param groupId 消费组
* @return Map[car-1 , car-2 , Long]
**/

private def readOffset(topic: String, groupId: String): Map[TopicPartition, Long] = {
val topicPartitionMap = collection.mutable.HashMap.empty[TopicPartition, Long]
//去zk上拿topic和分区信息
val topicAndPartitionMaps: mutable.Map[String, Seq[Int]] = zkUtils.getPartitionsForTopics(Seq(topic))
val groupId = kafkaParams("group.id").toString
val redisKey = topic + "|" + groupId
topicAndPartitionMaps.foreach(topicPartitions =>{
val zkGroupTopicsDirs: ZKGroupTopicDirs = new ZKGroupTopicDirs(groupId , topicPartitions._1)
topicPartitions._2.foreach(partition => {
//迭代分区
val map: util.Map[String, String] = jedis.hgetAll(redisKey)
val offsetMap: mutable.Map[String, String] = mapAsScalaMap(map)
if (offsetMap != null && offsetMap.size != 0) {
logger.error("groupId:"+groupId+"获取到redis的偏移量数据")
topicPartitionMap.put(new TopicPartition(topicPartitions._1, Integer.valueOf(partition)), offsetMap(partition.toString).toLong)
}
else {
logger.error("程序第一次启动,redis还未存储，获取kafka的偏移量")
val consumer = new KafkaConsumer[String, Object](kafkaParams)
val topicCollection = List(new TopicPartition(topicPartitions._1 , partition))
consumer.assign(topicCollection)
val avaliableOffset: Long = consumer.beginningOffsets(topicCollection).values().head
consumer.close()
topicPartitionMap.put(new TopicPartition(topicPartitions._1 , Integer.valueOf(partition)) , avaliableOffset)
}
})
}
)

//currentoffset 、 earliestoffset leatestOffset
//cur < ear || cur > leaty ==> 矫正--> ear
//TODO 矫正
val earliestOffsets = getEarliestOffsets(kafkaParams, topic)
val topics = List(topic)
val latestOffsets = getLatestOffsets(kafkaParams, topics)
for ((k, v) <- topicPartitionMap) {
val current = v
val earliest = earliestOffsets.get(k).get
val latest = latestOffsets.get(k).get
if (current < earliest || current > latest) {
topicPartitionMap.put(k, earliest)
}
}
topicPartitionMap.toMap
}

/**
* 获取最早的偏移量
*
* @param kafkaParams
* @param topics
* @return
*/
private def getEarliestOffsets(kafkaParams: Map[String, Object], topic: String) = {
val newKafkaParams = mutable.Map[String, Object]()
newKafkaParams ++= kafkaParams
newKafkaParams.put(ConsumerConfig.AUTO_OFFSET_RESET_CONFIG, "earliest")
//kafka api
val consumer = new KafkaConsumer(kafkaParams)
//订阅
val topics = Seq[String](topic)
consumer.subscribe(topics)
val noOffsetForPartitionExceptionSet: mutable.Set[Nothing] = mutable.Set()
try {
consumer.poll(0)
} catch {
case e: NoOffsetForPartitionException =>
// noOffsetForPartitionExceptionSet.add(e.partition())
//邮件报警
}
//获取分区信息
val topicp = consumer.assignment().toSet
//暂定消费
consumer.pause(topicp)
//从头开始
consumer.seekToBeginning(topicp)
val toMap = topicp.map(line => line -> consumer.position(line)).toMap
val earliestOffsetMap = toMap
consumer.unsubscribe()
consumer.close()
earliestOffsetMap
}

private def getLatestOffsets(kafkaParams: Map[String, Object], topic: Seq[String]) = {
val newKafkaParams = mutable.Map[String, Object]()
newKafkaParams ++= kafkaParams
newKafkaParams.put(ConsumerConfig.AUTO_OFFSET_RESET_CONFIG, "latest")

//kafka api
val consumer = new KafkaConsumer[String, Array[Byte]](newKafkaParams)
//订阅
consumer.subscribe(topic)
val noOffsetForPartitionExceptionSet = mutable.Set()
try {
consumer.poll(0)
} catch {
case e: NoOffsetForPartitionException =>
// noOffsetForPartitionExceptionSet.add(e.partition())
//邮件报警
}
//获取分区信息
val topicp = consumer.assignment().toSet
//暂定消费
consumer.pause(topicp)
//从尾开始
consumer.seekToEnd(topicp)
val toMap: Map[TopicPartition, Long] = topicp.map(line => line -> consumer.position(line)).toMap
val earliestOffsetMap = toMap
consumer.unsubscribe()
consumer.close()
earliestOffsetMap
}

def persistOffset[K, V](rdd: RDD[ConsumerRecord[K, V]], storeOffset: Boolean = true, topic: String) = {
val groupId = kafkaParams("group.id").toString
val offsetRanges: Array[OffsetRange] = rdd.asInstanceOf[HasOffsetRanges].offsetRanges
offsetRanges.foreach(offsetRange => {
val redisKey = topic + "|" + groupId
val data = if (storeOffset) offsetRange.untilOffset else offsetRange.fromOffset
jedis.hset(redisKey, offsetRange.partition.toString, data.toString)
println("topic:" + offsetRange.topic + "分区:" + offsetRange.partition + "开始消费" + offsetRange.fromOffset + "消费到" + offsetRange.untilOffset + "共计" + offsetRange.count())
})

}

}

object KafkaManagerByRedis {
def main(args: Array[String]): Unit = {
val kafkaParams = Map[String, Object](
"bootstrap.servers" -> "192.168.121.12:9092,192.168.121.12:9093",
"key.deserializer" -> classOf[StringDeserializer],
"value.deserializer" -> classOf[StringDeserializer],
"group.id" -> "test1",
"auto.offset.reset" -> ("earliest "),
"enable.auto.commit" -> (false: java.lang.Boolean) //禁用自动提交Offset，否则可能没正常消费完就提交了，造成数据错误
)
val zkServer=""
val kafkama = new KafkaManagerByRedis(zkServer,kafkaParams)
kafkama.getEarliestOffsets(kafkaParams, "cheng_du_gps_topic")
.foreach(m => println(m._1.topic(), m._1.partition(), m._2))

kafkama.getLatestOffsets(kafkaParams, List("cheng_du_gps_topic"))
.foreach(m => println(m._1.topic(), m._1.partition(), m._2))
}
}

巴特西

kafka-spark偏移量提交至redis kafka1.0版本

最新文章

热门文章