[ 
https://issues.apache.org/jira/browse/FLINK-7201?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16091053#comment-16091053
 ] 

ASF GitHub Bot commented on FLINK-7201:
---------------------------------------

Github user XuPingyong commented on the issue:

    https://github.com/apache/flink/pull/4347
  
    @StephanEwen ,  rpcService of ResourceManager executes with only one single 
thread, so there is no conflicts when resourcemanager is in service. When 
resourceManager is shutdown by the other thread, the rpcService had better stop 
first.
    



> ConcurrentModificationException in JobLeaderIdService
> -----------------------------------------------------
>
>                 Key: FLINK-7201
>                 URL: https://issues.apache.org/jira/browse/FLINK-7201
>             Project: Flink
>          Issue Type: Bug
>          Components: JobManager
>            Reporter: Xu Pingyong
>            Assignee: Xu Pingyong
>              Labels: flip-6
>
> {code:java}
>  java.util.ConcurrentModificationException: null
>       at java.util.HashMap$HashIterator.nextEntry(HashMap.java:922)
>       at java.util.HashMap$ValueIterator.next(HashMap.java:950)
>       at 
> org.apache.flink.runtime.resourcemanager.JobLeaderIdService.clear(JobLeaderIdService.java:114)
>       at 
> org.apache.flink.runtime.resourcemanager.JobLeaderIdService.stop(JobLeaderIdService.java:92)
>       at 
> org.apache.flink.runtime.resourcemanager.ResourceManager.shutDown(ResourceManager.java:200)
>       at 
> org.apache.flink.runtime.resourcemanager.ResourceManagerRunner.shutDownInternally(ResourceManagerRunner.java:102)
>       at 
> org.apache.flink.runtime.resourcemanager.ResourceManagerRunner.shutDown(ResourceManagerRunner.java:97)
>       at 
> org.apache.flink.runtime.minicluster.MiniCluster.shutdownInternally(MiniCluster.java:329)
>       at 
> org.apache.flink.runtime.minicluster.MiniCluster.shutdown(MiniCluster.java:297)
>       at 
> org.apache.flink.runtime.minicluster.MiniClusterITCase.runJobWithMultipleJobManagers(MiniClusterITCase.java:85)
> {code}
> Because the jobLeaderIdService stops before the rpcService when shutdown the 
> resourceManager, jobLeaderIdService has a risk of thread-unsafe.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Reply via email to