question Support for external shuffle services in Support Questions

Support for external shuffle services

dbrys — Wed, 24 May 2023 09:27:11 GMT

Hi, does anyone knows whether there are any restrictions on using external shuffle services on CDP?

For instance, are the following external shuffle services supported?:

- uber rss

- apache uniffle

Re: Support for external shuffle services

DianaTorres — Wed, 24 May 2023 15:33:22 GMT

@dbrys Welcome to the Cloudera Community!

To help you get the best possible solution, I have tagged our Spark experts @Bharati and @jagadeesan who may be able to assist you further.

Please keep us updated on your post, and we hope you find a satisfactory solution to your query.

Re: Support for external shuffle services

Bharati — Wed, 24 May 2023 18:21:08 GMT

At the moment both uber rss and apache uniffle are not supported in CDP.

Dynamic resource allocation requires an external shuffle service that runs on each worker node as an auxiliary service of NodeManager. This service is started automatically; no further steps are needed. spark.shuffle.service.enabled=true enables the external shuffle service. The external shuffle service preserves shuffle files written by executors so that the executors can be deallocated without losing work. Must be enabled if dynamic allocation is enabled.

Re: Support for external shuffle services

dbrys — Tue, 30 May 2023 06:28:40 GMT

thanks for the information, we're running on yarn so I guess the steps mentioned in https://spark.apache.org/docs/3.0.0-preview2/running-on-yarn.html#configuring-the-external-shuffle-service must be executed to configure the external shuffle service on each worker node. Are there any plans to support any other external shuffle service like uber rss or apache uniffle in the future?