gearpump-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (GEARPUMP-285) AppManager shut down executor time out
Date Thu, 09 Mar 2017 12:23:37 GMT

    [ https://issues.apache.org/jira/browse/GEARPUMP-285?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15902970#comment-15902970
] 

ASF GitHub Bot commented on GEARPUMP-285:
-----------------------------------------

Github user manuzhang commented on a diff in the pull request:

    https://github.com/apache/incubator-gearpump/pull/169#discussion_r105150887
  
    --- Diff: core/src/main/scala/org/apache/gearpump/cluster/master/AppManager.scala ---
    @@ -231,35 +231,38 @@ private[cluster] class AppManager(kvService: ActorRef, launcher:
AppMasterLaunch
           timeStamp: TimeStamp, error: Throwable): Unit = {
         applicationRegistry.get(appId) match {
           case Some(appRuntimeInfo) =>
    -        var updatedStatus: ApplicationRuntimeInfo = null
    -        LOG.info(s"Application $appId change to ${newStatus.toString} at $timeStamp")
    -        newStatus match {
    -          case ApplicationStatus.ACTIVE =>
    -            updatedStatus = appRuntimeInfo.onAppMasterActivated(timeStamp)
    -            sender ! AppMasterActivated(appId)
    -          case succeeded@ApplicationStatus.SUCCEEDED =>
    -            killAppMaster(appId, appRuntimeInfo.worker)
    -            updatedStatus = appRuntimeInfo.onFinalStatus(timeStamp, succeeded)
    -            appResultListeners.getOrElse(appId, List.empty).foreach{ client =>
    -              client ! ApplicationSucceeded(appId)
    -            }
    -          case failed@ApplicationStatus.FAILED =>
    -            killAppMaster(appId, appRuntimeInfo.worker)
    -            updatedStatus = appRuntimeInfo.onFinalStatus(timeStamp, failed)
    -            appResultListeners.getOrElse(appId, List.empty).foreach{ client =>
    -              client ! ApplicationFailed(appId, error)
    -            }
    -          case terminated@ApplicationStatus.TERMINATED =>
    -            updatedStatus = appRuntimeInfo.onFinalStatus(timeStamp, terminated)
    -          case status =>
    -            LOG.error(s"App $appId should not change it's status to $status")
    -        }
    +        // A dead application should not be able to update status
    +        if (!appRuntimeInfo.status.isInstanceOf[ApplicationTerminalStatus]) {
    --- End diff --
    
    Do we want to update status if not changed ? Can there be other "false alarm"s ?


> AppManager shut down executor time out
> --------------------------------------
>
>                 Key: GEARPUMP-285
>                 URL: https://issues.apache.org/jira/browse/GEARPUMP-285
>             Project: Apache Gearpump
>          Issue Type: Bug
>    Affects Versions: 0.8.2
>            Reporter: Manu Zhang
>            Assignee: Huafeng Wang
>            Priority: Minor
>
> When running Beam tests for about 10 seconds with {{EmbeddedCluster}}, I saw the "Shut
down executor time out" error log which should be trigger only after 30 seconds. 



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Mime
View raw message