[jira] [Created] (FLINK-22574) Adaptive Scheduler: Can not cancel restarting job

[Shang Yuanchun (Jira)]

Robert Metzger created FLINK-22574:
--------------------------------------

             Summary: Adaptive Scheduler: Can not cancel restarting job
                 Key: FLINK-22574
                 URL: https://issues.apache.org/jira/browse/FLINK-22574
             Project: Flink
          Issue Type: Bug
          Components: Runtime / Coordination
    Affects Versions: 1.13.0
            Reporter: Robert Metzger
             Fix For: 1.13.1


I have a job in state RESTARTING. When I now issue a cancel RPC call, I get the following exception:

{code}
org.apache.flink.runtime.rest.handler.RestHandlerException: Job cancellation failed: Cancellation failed.
 at org.apache.flink.runtime.rest.handler.job.JobCancellationHandler.lambda$handleRequest$0(JobCancellationHandler.java:127)
 at java.util.concurrent.CompletableFuture.uniHandle(CompletableFuture.java:836)
 at java.util.concurrent.CompletableFuture$UniHandle.tryFire(CompletableFuture.java:811)
 at java.util.concurrent.CompletableFuture.postComplete(CompletableFuture.java:488)
 at java.util.concurrent.CompletableFuture.completeExceptionally(CompletableFuture.java:1990)
 at org.apache.flink.runtime.rpc.akka.AkkaInvocationHandler.lambda$invokeRpc$0(AkkaInvocationHandler.java:234)
 at java.util.concurrent.CompletableFuture.uniWhenComplete(CompletableFuture.java:774)
 at java.util.concurrent.CompletableFuture$UniWhenComplete.tryFire(CompletableFuture.java:750)
 at java.util.concurrent.CompletableFuture.postComplete(CompletableFuture.java:488)
 at java.util.concurrent.CompletableFuture.completeExceptionally(CompletableFuture.java:1990)
 at org.apache.flink.runtime.concurrent.FutureUtils$1.onComplete(FutureUtils.java:1079)
 at akka.dispatch.OnComplete.internal(Future.scala:263)
 at akka.dispatch.OnComplete.internal(Future.scala:261)
 at akka.dispatch.japi$CallbackBridge.apply(Future.scala:191)
 at akka.dispatch.japi$CallbackBridge.apply(Future.scala:188)
 at scala.concurrent.impl.CallbackRunnable.run(Promise.scala:60)
 at org.apache.flink.runtime.concurrent.Executors$DirectExecutionContext.execute(Executors.java:73)
 at scala.concurrent.impl.CallbackRunnable.executeWithValue(Promise.scala:68)
 at scala.concurrent.impl.Promise$DefaultPromise.$anonfun$tryComplete$1(Promise.scala:284)
 at scala.concurrent.impl.Promise$DefaultPromise.$anonfun$tryComplete$1$adapted(Promise.scala:284)
 at scala.concurrent.impl.Promise$DefaultPromise.tryComplete(Promise.scala:284)
 at akka.pattern.PromiseActorRef.$bang(AskSupport.scala:573)
 at akka.pattern.PipeToSupport$PipeableFuture$$anonfun$pipeTo$1.applyOrElse(PipeToSupport.scala:23)
 at akka.pattern.PipeToSupport$PipeableFuture$$anonfun$pipeTo$1.applyOrElse(PipeToSupport.scala:21)
 at scala.concurrent.Future.$anonfun$andThen$1(Future.scala:532)
 at scala.concurrent.impl.Promise.liftedTree1$1(Promise.scala:29)
 at scala.concurrent.impl.Promise.$anonfun$transform$1(Promise.scala:29)
 at scala.concurrent.impl.CallbackRunnable.run(Promise.scala:60)
 at akka.dispatch.BatchingExecutor$AbstractBatch.processBatch(BatchingExecutor.scala:55)
 at akka.dispatch.BatchingExecutor$BlockableBatch.$anonfun$run$1(BatchingExecutor.scala:91)
 at scala.runtime.java8.JFunction0$mcV$sp.apply(JFunction0$mcV$sp.java:12)
 at scala.concurrent.BlockContext$.withBlockContext(BlockContext.scala:81)
 at akka.dispatch.BatchingExecutor$BlockableBatch.run(BatchingExecutor.scala:91)
 at akka.dispatch.TaskInvocation.run(AbstractDispatcher.scala:40)
 at akka.dispatch.ForkJoinExecutorConfigurator$AkkaForkJoinTask.exec(ForkJoinExecutorConfigurator.scala:44)
 at akka.dispatch.forkjoin.ForkJoinTask.doExec(ForkJoinTask.java:260)
 at akka.dispatch.forkjoin.ForkJoinPool$WorkQueue.runTask(ForkJoinPool.java:1339)
 at akka.dispatch.forkjoin.ForkJoinPool.runWorker(ForkJoinPool.java:1979)
 at akka.dispatch.forkjoin.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:107) Caused by: org.apache.flink.util.FlinkException: Cancellation failed.
 at org.apache.flink.runtime.jobmaster.JobMasterServiceLeadershipRunner.lambda$cancel$3(JobMasterServiceLeadershipRunner.java:197)
 at java.util.concurrent.CompletableFuture.uniExceptionally(CompletableFuture.java:884)
 at java.util.concurrent.CompletableFuture$UniExceptionally.tryFire(CompletableFuture.java:866)
 at java.util.concurrent.CompletableFuture.postComplete(CompletableFuture.java:488)
 at java.util.concurrent.CompletableFuture.completeExceptionally(CompletableFuture.java:1990)
 at org.apache.flink.runtime.rpc.akka.AkkaInvocationHandler.lambda$invokeRpc$0(AkkaInvocationHandler.java:234)
 at java.util.concurrent.CompletableFuture.uniWhenComplete(CompletableFuture.java:774)
 at java.util.concurrent.CompletableFuture$UniWhenComplete.tryFire(CompletableFuture.java:750)
 at java.util.concurrent.CompletableFuture.postComplete(CompletableFuture.java:488)
 at java.util.concurrent.CompletableFuture.completeExceptionally(CompletableFuture.java:1990)
 at org.apache.flink.runtime.concurrent.FutureUtils$1.onComplete(FutureUtils.java:1079)
 at akka.dispatch.OnComplete.internal(Future.scala:263)
 at akka.dispatch.OnComplete.internal(Future.scala:261)
 at akka.dispatch.japi$CallbackBridge.apply(Future.scala:191)
 at akka.dispatch.japi$CallbackBridge.apply(Future.scala:188)
 at scala.concurrent.impl.CallbackRunnable.run(Promise.scala:60)
 at org.apache.flink.runtime.concurrent.Executors$DirectExecutionContext.execute(Executors.java:73)
 at scala.concurrent.impl.CallbackRunnable.executeWithValue(Promise.scala:68)
 at scala.concurrent.impl.Promise$DefaultPromise.$anonfun$tryComplete$1(Promise.scala:284)
 at scala.concurrent.impl.Promise$DefaultPromise.$anonfun$tryComplete$1$adapted(Promise.scala:284)
 at scala.concurrent.impl.Promise$DefaultPromise.tryComplete(Promise.scala:284)
 at akka.pattern.PromiseActorRef.$bang(AskSupport.scala:573)
 at akka.actor.ActorRef.tell(ActorRef.scala:126)
 at org.apache.flink.runtime.rpc.akka.AkkaRpcActor.handleRpcInvocation(AkkaRpcActor.java:311)
 at org.apache.flink.runtime.rpc.akka.AkkaRpcActor.handleRpcMessage(AkkaRpcActor.java:212)
 at org.apache.flink.runtime.rpc.akka.FencedAkkaRpcActor.handleRpcMessage(FencedAkkaRpcActor.java:77)
 at org.apache.flink.runtime.rpc.akka.AkkaRpcActor.handleMessage(AkkaRpcActor.java:158)
 at akka.japi.pf.UnitCaseStatement.apply(CaseStatements.scala:26)
 at akka.japi.pf.UnitCaseStatement.apply(CaseStatements.scala:21)
 at scala.PartialFunction.applyOrElse(PartialFunction.scala:123)
 at scala.PartialFunction.applyOrElse$(PartialFunction.scala:122)
 at akka.japi.pf.UnitCaseStatement.applyOrElse(CaseStatements.scala:21)
 at scala.PartialFunction$OrElse.applyOrElse(PartialFunction.scala:171)
 at scala.PartialFunction$OrElse.applyOrElse(PartialFunction.scala:172)
 at scala.PartialFunction$OrElse.applyOrElse(PartialFunction.scala:172)
 at akka.actor.Actor.aroundReceive(Actor.scala:517)
 at akka.actor.Actor.aroundReceive$(Actor.scala:515)
 at akka.actor.AbstractActor.aroundReceive(AbstractActor.scala:225)
 at akka.actor.ActorCell.receiveMessage(ActorCell.scala:592)
 at akka.actor.ActorCell.invoke(ActorCell.scala:561)
 at akka.dispatch.Mailbox.processMailbox(Mailbox.scala:258)
 at akka.dispatch.Mailbox.run(Mailbox.scala:225)
 at akka.dispatch.Mailbox.exec(Mailbox.scala:235) ... 4 more Caused by: java.lang.IllegalStateException: Assuming running execution graph
 at org.apache.flink.util.Preconditions.checkState(Preconditions.java:193)
 at org.apache.flink.runtime.scheduler.adaptive.StateWithExecutionGraph.(StateWithExecutionGraph.java:94)
 at org.apache.flink.runtime.scheduler.adaptive.Canceling.(Canceling.java:41)
 at org.apache.flink.runtime.scheduler.adaptive.Canceling$Factory.getState(Canceling.java:98)
 at org.apache.flink.runtime.scheduler.adaptive.Canceling$Factory.getState(Canceling.java:72)
 at org.apache.flink.runtime.scheduler.adaptive.AdaptiveScheduler.transitionToState(AdaptiveScheduler.java:1128)
 at org.apache.flink.runtime.scheduler.adaptive.AdaptiveScheduler.goToCanceling(AdaptiveScheduler.java:802)
 at org.apache.flink.runtime.scheduler.adaptive.Restarting.cancel(Restarting.java:74)
 at org.apache.flink.runtime.scheduler.adaptive.AdaptiveScheduler.cancel(AdaptiveScheduler.java:453)
 at org.apache.flink.runtime.jobmaster.JobMaster.cancel(JobMaster.java:417)
 at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
 at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
 at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
 at java.lang.reflect.Method.invoke(Method.java:498)
 at org.apache.flink.runtime.rpc.akka.AkkaRpcActor.handleRpcInvocation(AkkaRpcActor.java:305) ... 23 more
{code}
(Sorry for the poor stack trace formatting, I coped the stack trace from the UI)

It seems that the Canceling state assumes we only transition into it from Executing (ExecutionGraph = RUNNING).

In my opinion a job should be cancellable at all times, for example when a job is stuck in a restart loop, cancelling is the only way out (unless retries are exhausted).



--
This message was sent by Atlassian Jira
(v8.3.4#803005)