[
https://issues.apache.org/jira/browse/YARN-6620?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16169566#comment-16169566
]
Zhankun Tang commented on YARN-6620:
------------------------------------
[~wangda], Thanks for the patch!
Now we have defined the resource plugin framework in NM which manages various
resources that implemented ResourcePlugin and ResourceHandler interface easily.
It also provide a good way for resource plugin to update NodeStatus which
provide possibility for FPGA IP update thru node labels/attributes.
Two questions/comments from me:
1. Current GPUResourceAllocator is not got from LocalResourceAllocators but
created in GpuResourceHandlerImpl directly. Is this intended?
2. The GpuResourceHandler get container's requested GPU from an environment key
"REQUESTED_GPU_NUM". So in fact, there's no need to define the allowed GPU
resource in "node-resouce.xml". This is not an issue at present, but may brings
a potential limitation that the end-user cannot declare different type of GPU
vendor resource. For instance, if different vendors' GPU cards are installed in
the cluster, how can a user distinguish them? thru node attributes?
> [YARN-6223] NM Java side code changes to support isolate GPU devices by using
> CGroups
> -------------------------------------------------------------------------------------
>
> Key: YARN-6620
> URL: https://issues.apache.org/jira/browse/YARN-6620
> Project: Hadoop YARN
> Issue Type: Sub-task
> Reporter: Wangda Tan
> Assignee: Wangda Tan
> Attachments: YARN-6620.001.patch, YARN-6620.002.patch,
> YARN-6620.003.patch, YARN-6620.004.patch, YARN-6620.005.patch,
> YARN-6620.006-WIP.patch
>
>
> This JIRA plan to add support of:
> 1) GPU configuration for NodeManagers
> 2) Isolation in CGroups. (Java side).
> 3) NM restart and recovery allocated GPU devices
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
