You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@mesos.apache.org by "Kevin Klues (JIRA)" <ji...@apache.org> on 2016/04/22 23:41:13 UTC

[jira] [Commented] (MESOS-5255) Add support for GPU usage in metrics endpoint

    [ https://issues.apache.org/jira/browse/MESOS-5255?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15254748#comment-15254748 ] 

Kevin Klues commented on MESOS-5255:
------------------------------------

This will involve changes to the ResourceStatistics protobuf as well:
{noformat}
include/mesos/mesos.proto
include/mesos/v1/mesos.proto
{noformat}

It will also involve changes to the webui to report this information (in file):
{noformat}
src/webui/master/static/slave_executor.html
{noformat}

Look for the line below and mimic what is done for CPUs:
{noformat}
            <td>GPUs</td>
            <td class="text-right">
              N/A
            </td>
{noformat}

Also in:
{noformat}
src/webui/master/static/slave_framework.html
{noformat}

{noformat}
          <tr>
            <td>GPUs</td>
            <td class="text-right">
              N/A
            </td>
            <td class="text-right">{{framework.gpus | number}}</td>
          </tr>
{noformat}

> Add support for GPU usage in metrics endpoint
> ---------------------------------------------
>
>                 Key: MESOS-5255
>                 URL: https://issues.apache.org/jira/browse/MESOS-5255
>             Project: Mesos
>          Issue Type: Task
>            Reporter: Kevin Klues
>            Assignee: Kevin Klues
>              Labels: gpu
>
> Currently the usage callback in the Nvidia GPU isolator is unimplemented:
> {noformat}
> src/slave/containerizer/mesos/isolators/cgroups/devices/gpus/nvidia.cpp
> {noformat}
> It should use functionality from NVML to gather the current GPU usage and add it to a ResourceStatistics object. It is still an open question as to exactly what information we want to expose here (power, memory consumption, current load, etc.). Whatever we decide on should be standard across different GPU types, different GPU vendors, etc.



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