You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Marcelo Masiero Vanzin (Jira)" <ji...@apache.org> on 2019/11/26 00:16:00 UTC
[jira] [Commented] (SPARK-30033) Use Spark plugin support to manage
shuffle plugin lifecycle
[ https://issues.apache.org/jira/browse/SPARK-30033?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16981997#comment-16981997 ]
Marcelo Masiero Vanzin commented on SPARK-30033:
------------------------------------------------
[~mcheah] [~yifeih]
I had plans to do this at some point, but after taking a look at your S3 prototype it seems you guys already have a use case for the extra functionality Spark already has. I'll send a PR soon with my current implementation.
> Use Spark plugin support to manage shuffle plugin lifecycle
> -----------------------------------------------------------
>
> Key: SPARK-30033
> URL: https://issues.apache.org/jira/browse/SPARK-30033
> Project: Spark
> Issue Type: Sub-task
> Components: Spark Core
> Affects Versions: 3.0.0
> Reporter: Marcelo Masiero Vanzin
> Priority: Major
>
> SPARK-29396 added support for Spark plugins that have driver and executor components. It provides lifecycle APIs (initialization / shutdown) and also some extra features like metric registration and easy access to Spark's RPC system.
> We could use those APIs as the base for the shuffle-related plugins, so that they have easy access to the extra functionality if they need to.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org