warrenzhu25 commented on PR #37822:
URL: https://github.com/apache/spark/pull/37822#issuecomment-1270746315

   > Whats the intent of recommissioning? I don't super get it normally 
decommissioning means the worker is going away is this for the situation where 
(for example) planned maintance is cancelled?
   
   @holdenk Thanks for looking at this. We're doing manually scaling up and 
down workers based on spark metric. Manually means we're using master rest api 
to kill workers. Let's say we have 10 nodes to be decommissioned. 9 nodes 
finished quickly, but the last one took long as it might have more shuffle data 
to migrate or some long tail tasks still running. So we want to cancel the long 
tail node decommission to avoid shuffle data loss or task failure as during 
decommission it can't run new tasks it's waste of resource.


-- 
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