java.lang.reflect.InvocationTargetException at shade.com.datastax.spark.connector.google.common.base.Throwables.propagate(Throwables.java160)

org.apache.spark.SparkException: Job aborted due to stage failure: Task 1 in stage 29.1 failed 4 times, most recent failure: Lost task 1.3 in stage 29.1 (TID 466, magnesium, executor 4): java.lang.RuntimeException: java.lang.reflect.InvocationTargetException
at shade.com.datastax.spark.connector.google.common.base.Throwables.propagate(Throwables.java:160)
at com.datastax.driver.core.NettyUtil.newEventLoopGroupInstance(NettyUtil.java:136)
at com.datastax.driver.core.NettyOptions.eventLoopGroup(NettyOptions.java:99)
at com.datastax.driver.core.Connection$Factory.

我是在SparkStreaming查詢Cassandra時遇到這個報錯的。

dataFrame.foreachPartition { part =>
  val poolingOptions = new PoolingOptions
  poolingOptions
    .setCoreConnectionsPerHost(HostDistance.LOCAL, 4)
    .setMaxConnectionsPerHost(HostDistance.LOCAL, 10)
  val cluster = Cluster
    .builder
    .addContactPoints("localhost")
    .withCredentials("cassandra", "cassandra")
    .withPoolingOptions(poolingOptions)
    .build
  val session = cluster.connect("keyspace")
  part.foreach { item =>
    // 業務邏輯
  }
  cluster.close()
  session.close()
}

每個批次中，首先檢查cluster和session，是否都close，沒有close會報這個錯誤。

若還沒有解決，需要檢查netty的版本。

推薦在IDEA中安裝Maven Helper插件。然後把衝突的低版本netty相關的依賴刪掉即可。