You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Stephan Ewen (Jira)" <ji...@apache.org> on 2019/11/21 15:14:00 UTC
[jira] [Resolved] (FLINK-14735) Improve batch schedule check input
consumable performance
[ https://issues.apache.org/jira/browse/FLINK-14735?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Stephan Ewen resolved FLINK-14735.
----------------------------------
Resolution: Fixed
Fixed in 1.10 in 93dfdd05a84f933473c7b22437e12c03239f9462 and d190308a16011384fac53484bb18cf61afd8cdc9
> Improve batch schedule check input consumable performance
> ---------------------------------------------------------
>
> Key: FLINK-14735
> URL: https://issues.apache.org/jira/browse/FLINK-14735
> Project: Flink
> Issue Type: Sub-task
> Components: Runtime / Task
> Reporter: Jingsong Lee
> Assignee: Zhu Zhu
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.10.0
>
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Now if we launch batch job with 1000+ parallelism:
> Even if we set the akka timeout of 2 minutes, the heartbeat is likely to timeout.
> JobMaster is buzy:
> {code:java}
> java.lang.Thread.State: RUNNABLE
> at java.util.stream.ReferencePipeline$3$1.accept(ReferencePipeline.java:193)
> at java.util.Spliterators$ArraySpliterator.tryAdvance(Spliterators.java:958)
> at java.util.stream.ReferencePipeline.forEachWithCancel(ReferencePipeline.java:126)
> at java.util.stream.AbstractPipeline.copyIntoWithCancel(AbstractPipeline.java:498)
> at java.util.stream.AbstractPipeline.copyInto(AbstractPipeline.java:485)
> at java.util.stream.AbstractPipeline.wrapAndCopyInto(AbstractPipeline.java:471)
> at java.util.stream.MatchOps$MatchOp.evaluateSequential(MatchOps.java:230)
> at java.util.stream.MatchOps$MatchOp.evaluateSequential(MatchOps.java:196)
> at java.util.stream.AbstractPipeline.evaluate(AbstractPipeline.java:234)
> at java.util.stream.ReferencePipeline.anyMatch(ReferencePipeline.java:449)
> at org.apache.flink.runtime.executiongraph.ExecutionVertex.isInputConsumable(ExecutionVertex.java:824)
> at org.apache.flink.runtime.executiongraph.ExecutionVertex$$Lambda$257/564237119.test(Unknown Source)
> at java.util.stream.MatchOps$2MatchSink.accept(MatchOps.java:119)
> at java.util.stream.Streams$RangeIntSpliterator.tryAdvance(Streams.java:89)
> at java.util.stream.IntPipeline.forEachWithCancel(IntPipeline.java:162)
> at java.util.stream.AbstractPipeline.copyIntoWithCancel(AbstractPipeline.java:498)
> at java.util.stream.AbstractPipeline.copyInto(AbstractPipeline.java:485)
> at java.util.stream.AbstractPipeline.wrapAndCopyInto(AbstractPipeline.java:471)
> at java.util.stream.MatchOps$MatchOp.evaluateSequential(MatchOps.java:230)
> at java.util.stream.MatchOps$MatchOp.evaluateSequential(MatchOps.java:196)
> at java.util.stream.AbstractPipeline.evaluate(AbstractPipeline.java:234)
> at java.util.stream.IntPipeline.allMatch(IntPipeline.java:482)
> at org.apache.flink.runtime.executiongraph.ExecutionVertex.checkInputDependencyConstraints(ExecutionVertex.java:811)
> at org.apache.flink.runtime.executiongraph.Execution.scheduleOrUpdateConsumers(Execution.java:889)
> at org.apache.flink.runtime.executiongraph.Execution.markFinished(Execution.java:1074)
> at org.apache.flink.runtime.executiongraph.ExecutionGraph.updateStateInternal(ExecutionGraph.java:1597)
> at org.apache.flink.runtime.executiongraph.ExecutionGraph.updateState(ExecutionGraph.java:1570)
> at org.apache.flink.runtime.scheduler.SchedulerBase.updateTaskExecutionState(SchedulerBase.java:424)
> at org.apache.flink.runtime.jobmaster.JobMaster.updateTaskExecutionState(JobMaster.java:380)
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)