Junping Du commented on YARN-3304:

I would like to get appendix (v2) patch in not only because it fix an issue of 
existing patch but also it shows how collaborative we are to downstream 
projects, like Tez. 
>From what I see from discussions above, they already did what they can do for 
>Hadoop part - delivered a patch to make API public and wait for the commit. It 
>is us - a small group of people watching this JIRA earlier - decided to change 
>API (for more readable reason) at last minute of release. The whole story here 
>sounds very interesting and like a multi-threads race condition happen on real 
>world.  Through appendix patch, we get sync on the same page again.
So, please help to review and get fix in. Thanks!

> ResourceCalculatorProcessTree#getCpuUsagePercent default return value is 
> inconsistent with other getters
> --------------------------------------------------------------------------------------------------------
>                 Key: YARN-3304
>                 URL: https://issues.apache.org/jira/browse/YARN-3304
>             Project: Hadoop YARN
>          Issue Type: Bug
>          Components: nodemanager
>            Reporter: Junping Du
>            Assignee: Junping Du
>            Priority: Blocker
>             Fix For: 2.7.0
>         Attachments: YARN-3304-appendix-v2.patch, YARN-3304-appendix.patch, 
> YARN-3304-v2.patch, YARN-3304-v3.patch, YARN-3304-v4-boolean-way.patch, 
> YARN-3304-v4-negative-way-MR.patch, YARN-3304-v4-negtive-value-way.patch, 
> YARN-3304-v6-no-rename.patch, YARN-3304-v6-with-rename.patch, 
> YARN-3304-v7.patch, YARN-3304-v8.patch, YARN-3304.patch, yarn-3304-5.patch
> Per discussions in YARN-3296, getCpuUsagePercent() will return -1 for 
> unavailable case while other resource metrics are return 0 in the same case 
> which sounds inconsistent.

This message was sent by Atlassian JIRA

Reply via email to