2

I am trying to kill my spark-kafka streaming job from Spark UI. It is able to kill the application but the driver is still running.

Can anyone help me with this. I am good with my other streaming jobs. only one of the streaming jobs is giving this problem ever time.

I can't kill the driver through command or spark UI. Spark Master is alive.

Output i collected from logs is -

16/10/25 03:14:25 INFO BlockManagerMaster: Removed 0 successfully in removeExecutor
16/10/25 03:14:25 INFO SparkUI: Stopped Spark web UI at http://***:4040
16/10/25 03:14:25 INFO SparkDeploySchedulerBackend: Shutting down all executors
16/10/25 03:14:25 INFO SparkDeploySchedulerBackend: Asking each executor to shut down
16/10/25 03:14:35 INFO AppClient: Stop request to Master timed out; it may already be shut down.
16/10/25 03:14:35 INFO MapOutputTrackerMasterEndpoint: MapOutputTrackerMasterEndpoint stopped!
16/10/25 03:14:35 INFO MemoryStore: MemoryStore cleared
16/10/25 03:14:35 INFO BlockManager: BlockManager stopped
16/10/25 03:14:35 INFO BlockManagerMaster: BlockManagerMaster stopped
16/10/25 03:14:35 INFO OutputCommitCoordinator$OutputCommitCoordinatorEndpoint: OutputCommitCoordinator stopped!
16/10/25 03:14:35 INFO SparkContext: Successfully stopped SparkContext
16/10/25 03:14:35 ERROR Inbox: Ignoring error
org.apache.spark.SparkException: Exiting due to error from cluster scheduler: Master removed our application: KILLED
    at org.apache.spark.scheduler.TaskSchedulerImpl.error(TaskSchedulerImpl.scala:438)
    at org.apache.spark.scheduler.cluster.SparkDeploySchedulerBackend.dead(SparkDeploySchedulerBackend.scala:124)
    at org.apache.spark.deploy.client.AppClient$ClientEndpoint.markDead(AppClient.scala:264)
    at org.apache.spark.deploy.client.AppClient$ClientEndpoint$$anonfun$receive$1.applyOrElse(AppClient.scala:172)
    at org.apache.spark.rpc.netty.Inbox$$anonfun$process$1.apply$mcV$sp(Inbox.scala:116)
    at org.apache.spark.rpc.netty.Inbox.safelyCall(Inbox.scala:204)
    at org.apache.spark.rpc.netty.Inbox.process(Inbox.scala:100)
    at org.apache.spark.rpc.netty.Dispatcher$MessageLoop.run(Dispatcher.scala:215)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
    at java.lang.Thread.run(Thread.java:745)
16/10/25 03:14:35 WARN NettyRpcEnv: Ignored message: true
16/10/25 03:14:35 WARN AppClient$ClientEndpoint: Connection to master:7077 failed; waiting for master to reconnect...
16/10/25 03:14:35 WARN AppClient$ClientEndpoint: Connection to master:7077 failed; waiting for master to reconnect...
AKC
  • 953
  • 4
  • 17
  • 46

1 Answers1

0

Get the running driverId from spark UI, and hit the post rest call(spark master rest port like 6066) to kill the pipeline. I have tested it with spark 1.6.1

curl -X POST http://localhost:6066/v1/submissions/kill/driverId

Hope it helps...

Anupam Jain
  • 476
  • 2
  • 6