Naganarasimha G R commented on YARN-4176:

Thanks [~bibinchundatt] for working on this issue, few small nits remaining :
# {{nodeLabelsProvider}} and {{resyncInterval}} can be final
# {{lastNodeLabelSendMills}} needs to be set even in case validation fails this 
will avoid invalid labels to be validated for every heartbeat
# In TestNodeStatusUpdaterForLabels, 
{{dummyLabelsProviderRef.setNodeLabels(toNodeLabelSet("P1"));}} can be outside 
the for loop and {{assertNull("Null resourcelabels are expected", 
resourceTracker.labels);}} is not required in side the if check.
# Modify {{updatedLabelsSentToRM}} -> areLabelsSentToRM
# instead of logging on every resync in 
{{verifyRMHeartbeatResponseForNodeLabels}} we can either make it debug log or 
based on labels update/RM reject we can log. I prefer for the former approach.

Apart from this every thing else is fine and have tested the patch in the 
[~wangda] we have kept the sync interval as 2 mins so that when cluster 
nodelabels are added, NM HB can update latest labels fast, please share weather 
we need to increase this sync duration to a greater value like 10 mins. 

> Resync NM nodelabels with RM every x interval for distributed nodelabels
> ------------------------------------------------------------------------
>                 Key: YARN-4176
>                 URL: https://issues.apache.org/jira/browse/YARN-4176
>             Project: Hadoop YARN
>          Issue Type: Bug
>            Reporter: Bibin A Chundatt
>            Assignee: Bibin A Chundatt
>         Attachments: 0001-YARN-4176.patch, 0002-YARN-4176.patch, 
> 0003-YARN-4176.patch, 0004-YARN-4176.patch
> This JIRA is for handling the below set of issue
> # Distributed nodelabels after NM registered with RM if cluster nodelabels 
> are removed and added then NM doesnt resend labels in heartbeat again untils 
> any change in labels
> # NM registration failed with Nodelabels should resend labels again to RM 
> The above cases can be handled by  resync nodeLabels with RM every x interval
> # Add property {{yarn.nodemanager.node-labels.provider.resync-interval-ms}} 
> and  will resend nodelabels to RM based on config no matter what the 
> registration fails or success.

This message was sent by Atlassian JIRA

Reply via email to