SparkStream：4)foreachRDD詳解

轉載自：http://blog.csdn.net/jiangpeng59/article/details/53318761

 

foreachRDD通常用來把SparkStream運行得到的結果保存到外部系統比如HDFS、Mysql、Redis等等。了解下面的知識可以幫助我們避免很多誤區

 

誤區1：實例化外部連接對象的位置不正確，比如下面代碼

dstream.foreachRDD { rdd =>
  val connection = createNewConnection()  // executed at the driver
  rdd.foreach { record =>
    connection.send(record) // executed at the worker
  }
}  ​

其實例化的連接對象在driver中，然后通過序列化的方式發送到各個Worker，但實際上Connection的序列化通常是無法正確序列化的

 

誤區2：為每條記錄都創建一個連接對象

dstream.foreachRDD { rdd =>
  rdd.foreach { record =>
    val connection = createNewConnection()
    connection.send(record)
    connection.close()
  }
}  ​

雖然誤區1的問題得到了解決，但通常情況下，外部系統如mysql，其連接對象是非常可貴的，如果一條記錄就申請一個連接資源，系統性能會非常糟糕

 

然后，給出了一個比較好的方法，為每一個分區創建一個連接對象，其具體代碼如下

dstream.foreachRDD { rdd =>
  rdd.foreachPartition { partitionOfRecords =>
    val connection = createNewConnection()
    partitionOfRecords.foreach(record => connection.send(record))
    connection.close()
  }
}  ​

最后給出一個較優的方案，使用一個連接池來維護連接對象

dstream.foreachRDD { rdd =>
  rdd.foreachPartition { partitionOfRecords =>
    // ConnectionPool is a static, lazily initialized pool of connections
    val connection = ConnectionPool.getConnection()
    partitionOfRecords.foreach(record => connection.send(record))
    ConnectionPool.returnConnection(connection)  // return to the pool for future reuse
  }
}  ​

正如上面代碼闡述的，連接對象推薦是使用lazy關鍵字來修飾，用到的時候才去實例化

 

下面給出網上一段把SparkStream的結果保存到Mysql中的代碼示例

package spark.examples.streaming

import java.sql.{PreparedStatement, Connection, DriverManager}
import java.util.concurrent.atomic.AtomicInteger

import org.apache.spark.SparkConf
import org.apache.spark.streaming.{Seconds, StreamingContext}
import org.apache.spark.streaming._
import org.apache.spark.streaming.StreamingContext._

object SparkStreamingForPartition {
  def main(args: Array[String]) {
    val conf = new SparkConf().setAppName("NetCatWordCount")
    conf.setMaster("local[3]")
    val ssc = new StreamingContext(conf, Seconds(5))
    //The DStream is a collection of RDD, which makes the method foreachRDD reasonable
    val dstream = ssc.socketTextStream("192.168.26.140", 9999)
    dstream.foreachRDD(rdd => {
      //embedded function
      def func(records: Iterator[String]) {
        var conn: Connection = null
        var stmt: PreparedStatement = null
        try {
          val url = "jdbc:mysql://192.168.26.140:3306/person";
          val user = "root";
          val password = ""
          conn = DriverManager.getConnection(url, user, password)
          records.flatMap(_.split(" ")).foreach(word => {
            val sql = "insert into TBL_WORDS(word) values (?)";
            stmt = conn.prepareStatement(sql);
            stmt.setString(1, word)
            stmt.executeUpdate();
          })
        } catch {
          case e: Exception => e.printStackTrace()
        } finally {
          if (stmt != null) {
            stmt.close()
          }
          if (conn != null) {
            conn.close()
          }
        }
      }
      val repartitionedRDD = rdd.repartition(3)
      repartitionedRDD.foreachPartition(func)
    })
    ssc.start()
    ssc.awaitTermination()
  }
}  ​

注意的細節：

Dstream和RDD一樣是延遲執行，只有遇到action操作才會真正去計算。因此在Dstream的內部RDD必須包含Action操作才能是接受到的數據得到處理。即使代碼中包含foreachRDD,但在內部卻沒有action的RDD，SparkStream只會簡單地接受數據數據而不進行處理