ScrapCodes commented on pull request #29368:
URL: https://github.com/apache/spark/pull/29368#issuecomment-670346087


   I had say, it should be discussed, we would like to hear from the users. Do 
we have enough confidence, for example if it is already in use at some company 
X at a production scale. Would that give us enough confidence.
   
   It is possible, that after external shuffle service is developed, the API 
may need to change. But, since we do not expose anything in K8s as API, it 
should not be a problem. Even if we do, we can keep the APIs as experimental, 
instead of declaring that "Spark on K8s is experimental and not ready for 
production, until ..." 
   
   Today, I see "so many" people (AFAIK) still using standalone mode on K8s for 
deploying spark on k8s. It is true, we may be missing a feature or two. But, 
all of this should be discussed and we should have consensus about when to GA.


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
[email protected]



---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to