tgravescs commented on issue #25299: [SPARK-27651][Core] Avoid the network when 
shuffle blocks are fetched from the same host
URL: https://github.com/apache/spark/pull/25299#issuecomment-558202319
 
 
   I only skimmed this but overall looks good, like this approach.
   
   When fetching the host local blocks do we have any limits on # that can 
happen concurrently?  I didn't see anything config on just skimming it. If we 
have a lot of other executors on the same node we end up bypassing the remote 
fetch configs as to # in parallel so now we are doing more work at once here 
which could potentially cause more local disk activity on both the read and 
write sides.

----------------------------------------------------------------
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]


With regards,
Apache Git Services

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

Reply via email to