tgravescs commented on pull request #31896:
URL: https://github.com/apache/spark/pull/31896#issuecomment-804125940


   so I would like to see more design details on this. I get that this adds a 
plugin for spark submit, but there are lots of other code in here that does 
stuff specific to cluster manager.  It also seems to assume that that plugin 
supports various options. I think we need to clearly define what specifically 
it will and will not handle and make sure that is documented.
   
   Some basic things are just like does it support both client and cluster 
deploy modes, which options does it support, how does it deal with 
dependencies, how does it deal with security, does it work with PySpark, 
sparkR, spark-shell, spark-submit, etc.
   I think to do this properly we would need more APIs between these so it can 
properly figure out what to do.
   
   kind of minor but I think this should also be referred to as cluster manager 
plugin or something like that so we know what the plugin is for.
   
   if we are expecting users to implement why is this private[spark]?
   private[spark] trait SparkSubmitPlugin
   
   we would want this a developerapi unstable and made public.


-- 
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