Mailing-List: contact yarn-issues-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Date: Mon, 10 Jul 2017 11:50:00 +0000 (UTC)
From: "Bibin A Chundatt (JIRA)" <jira@apache.org>
To: yarn-issues@hadoop.apache.org
Message-ID: <JIRA.12970523.1463515951000.206527.1499687400146@Atlassian.JIRA>
In-Reply-To: <JIRA.12970523.1463515951000@Atlassian.JIRA>
References: <JIRA.12970523.1463515951000@Atlassian.JIRA> <JIRA.12970523.1463515951349@jira-lw-us.apache.org>
Subject: [jira] [Commented] (YARN-5103) With NM recovery enabled, restarting
 NM multiple times results in AM restart
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
archived-at: Mon, 10 Jul 2017 11:50:11 -0000


    [ https://issues.apache.org/jira/browse/YARN-5103?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16080212#comment-16080212 ] 

Bibin A Chundatt commented on YARN-5103:
----------------------------------------

Hi [~djp]
Even after this change the NM would still report container  as *LOST* rt??.
    int retCode = ExitCode.LOST.getExitCode();
{code}
    if (retCode != 0) {
      LOG.warn("Recovered container exited with a non-zero exit code "
          + retCode);
      this.dispatcher.getEventHandler().handle(new ContainerExitEvent(
          containerId,
          ContainerEventType.CONTAINER_EXITED_WITH_FAILURE, retCode,
          "Container exited with a non-zero exit code " + retCode));
      return retCode;
    }
{code}
 or am I missing something?

> With NM recovery enabled, restarting NM multiple times results in AM restart
> ----------------------------------------------------------------------------
>
>                 Key: YARN-5103
>                 URL: https://issues.apache.org/jira/browse/YARN-5103
>             Project: Hadoop YARN
>          Issue Type: Bug
>          Components: yarn
>            Reporter: Sumana Sathish
>            Assignee: Junping Du
>            Priority: Critical
>             Fix For: 2.8.0, 3.0.0-alpha1
>
>         Attachments: YARN-5103-demo.patch, YARN-5103.patch, YARN-5103-v2.patch
>
>
> AM is restarted when NM is restarted multiple times even though NM recovery is enabled.
> {Code:title=NM log on which AM attempt 1 was running }
>  ERROR launcher.RecoveredContainerLaunch (RecoveredContainerLaunch.java:call(88)) - Unable to recover container container_e12_1463043063682_0002_01_000001
> java.io.IOException: java.lang.InterruptedException
> 	at org.apache.hadoop.util.Shell.runCommand(Shell.java:579)
> 	at org.apache.hadoop.util.Shell.run(Shell.java:487)
> 	at org.apache.hadoop.util.Shell$ShellCommandExecutor.execute(Shell.java:753)
> 	at org.apache.hadoop.yarn.server.nodemanager.LinuxContainerExecutor.signalContainer(LinuxContainerExecutor.java:478)
> 	at org.apache.hadoop.yarn.server.nodemanager.LinuxContainerExecutor.isContainerProcessAlive(LinuxContainerExecutor.java:542)
> 	at org.apache.hadoop.yarn.server.nodemanager.ContainerExecutor.reacquireContainer(ContainerExecutor.java:185)
> 	at org.apache.hadoop.yarn.server.nodemanager.LinuxContainerExecutor.reacquireContainer(LinuxContainerExecutor.java:445)
> 	at org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.RecoveredContainerLaunch.call(RecoveredContainerLaunch.java:83)
> 	at org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.RecoveredContainerLaunch.call(RecoveredContainerLaunch.java:46)
> 	at java.util.concurrent.FutureTask.run(FutureTask.java:266)
> 	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> 	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> 	at java.lang.Thread.run(Thread.java:745)
> {Code}


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

---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org