[ 
https://issues.apache.org/jira/browse/YARN-3842?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14596756#comment-14596756
 ] 

Karthik Kambatla commented on YARN-3842:
----------------------------------------

Thanks for the quick turnaround on this, Robert. 

One nit-pick on the test: would the following be more concise? 

{code}
        if (retryCount < 5) {
          retryCount++;
          if (isExpectingNMNotYetReadyException) {
            containerManager.setBlockNewContainerRequests(true);
          } else {
            throw new java.net.ConnectException("start container exception");
          }
        } else {
          containerManager.setBlockNewContainerRequests(false);
        }
        return super.startContainers(requests);
{code}

> NM restarts could lead to app failures
> --------------------------------------
>
>                 Key: YARN-3842
>                 URL: https://issues.apache.org/jira/browse/YARN-3842
>             Project: Hadoop YARN
>          Issue Type: Bug
>    Affects Versions: 2.7.0
>            Reporter: Karthik Kambatla
>            Assignee: Robert Kanter
>            Priority: Critical
>         Attachments: MAPREDUCE-6409.001.patch, MAPREDUCE-6409.002.patch, 
> YARN-3842.001.patch
>
>
> Consider the following scenario:
> 1. RM assigns a container on node N to an app A.
> 2. Node N is restarted
> 3. A tries to launch container on node N.
> 3 could lead to an NMNotYetReadyException depending on whether NM N has 
> registered with the RM. In MR, this is considered a task attempt failure. A 
> few of these could lead to a task/job failure.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to