Sunil G updated YARN-4170:
    Attachment: 0007-YARN-4170.patch

{{TestCapacityScheduler.testPreemptionInfo}} failure is not related to this 
patch. Test case can run successfully in its local runs. However this is a 
known sporadic issue and YARN-3533 fixed it in {{MockRM#launchAM}}. But 
{{TestCapacityScheduler.testPreemptionInfo}} is using a local {{launchAM}} 
within TestCapacityScheduler which is similar to MockRM api and doesnt have the 
fix in YARN-3533. I feel we do not need this local method, rather 
{{MockRM#launchAM}} can be used instead.
I am fixing this also along with this new patch. Please suggest if needed to be 
tracked separately.

> AM need to be notified with priority in AllocateResponse 
> ---------------------------------------------------------
>                 Key: YARN-4170
>                 URL: https://issues.apache.org/jira/browse/YARN-4170
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: resourcemanager
>            Reporter: Sunil G
>            Assignee: Sunil G
>         Attachments: 0001-YARN-4170.patch, 0002-YARN-4170.patch, 
> 0003-YARN-4170.patch, 0004-YARN-4170.patch, 0005-YARN-4170-v2.patch, 
> 0006-YARN-4170.patch, 0007-YARN-4170.patch
> As discussed in MAPREDUCE-5870, Application Master need to be notified with 
> priority in Allocate heartbeat.  This will help AM to know the priority and 
> can update JobStatus when client asks. 

This message was sent by Atlassian JIRA

Reply via email to