Github user mccheah commented on the issue:
https://github.com/apache/spark/pull/21390
For Kubernetes, without a story around the external shuffle service, all
the scratch space used by executors will be cleaned up by Kubernetes itself.
When we want shuffle data to persist across executor restarts we'll have to
think about this matter more carefully.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
