warrenzhu25 commented on PR #41076:
URL: https://github.com/apache/spark/pull/41076#issuecomment-1537293952

   > Thank you for pinging me. Could you improve this proposal more, 
@warrenzhu25 ?
   > 
   > * This PR claims that data migration can hurt performance, but the code is 
applied in all cases including the case data migration (shuffle/rdd) is 
disabled. Moreover, in general, when the migration data size is small, the 
claim is invalid.
   > * In the same way, I can imagine this PR introduces another regression in 
terms of resource utilizations. For example, Spark Thrift Server can 
decommission all executors at the same time, but this configuration (< 0.1) may 
hurt the speed of scaling down.
   > * It would be great if we can have a clear documentation about the 
relationship between this and other decommission configs.
   
   `maxRatio` has default value 1, so it's same as current behavior. Users has 
full control of this ratio based on their judgement of storage migration size 
and impact.


-- 
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.

To unsubscribe, e-mail: [email protected]

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