hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Xuan Gong (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-479) NM retry behavior for connection to RM should be similar for lost heartbeats
Date Thu, 21 Mar 2013 04:45:40 GMT

    [ https://issues.apache.org/jira/browse/YARN-479?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13608632#comment-13608632
] 

Xuan Gong commented on YARN-479:
--------------------------------

Couple of comments on the latest one (479-2):
1. In the while(true) loop at NodeStatusUpdaterImpl : startStatusUpdater() :
     rmRetryCount ++ and response = resourceTracker.nodeHeartbeat(request).getHeartbeatResponse()
can be in the try block, others such as NodeStatus nodeStatus = getNodeStatus(), etc, I think
we can move them out of while(true) loop. We only consider losting heartBeatResponse.
2.please re-phrase the warning message and error message for more clarity - something along
the lines of did not get the heartbeat response ...
3. testNMRegistration may not be a good place to test the changes. You can re-write your own
ResourceTracker and NodeStatusUpdater to mimic the heartbeat response lose, and test your
code if it can handle properly. Take a look at the MyNodeStatusUpdater and MyResourceTracker
class, they can tell you how to do that.
                
> NM retry behavior for connection to RM should be similar for lost heartbeats
> ----------------------------------------------------------------------------
>
>                 Key: YARN-479
>                 URL: https://issues.apache.org/jira/browse/YARN-479
>             Project: Hadoop YARN
>          Issue Type: Bug
>            Reporter: Hitesh Shah
>            Assignee: jian he
>         Attachments: YARN-479.1.patch, YARN-479.2.patch
>
>
> Regardless of connection loss at the start or at an intermediate point, NM's retry behavior
to the RM should follow the same flow. 

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message