You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by GitBox <gi...@apache.org> on 2019/07/10 09:18:44 UTC

[GitHub] [flink] tillrohrmann commented on issue #9058: [FLINK-13166] Add support for batch slot requests to SlotPoolImpl

tillrohrmann commented on issue #9058: [FLINK-13166] Add support for batch slot requests to SlotPoolImpl
URL: https://github.com/apache/flink/pull/9058#issuecomment-509982669
 
 
   Hi @xintongsong, you are right that we don't retry slot requests from within the `SlotPool`. However, we do so in case of a failover.
   
   I'm actually not so sure whether the `SlotPool` should retry the requests. One could also argue that the `SlotManager` should simply not fail the slot requests if it thinks that it can eventually satisfy it. In both cases you need to define a timeout, though, and it just happens that in the current implementation we don't retry.
   
   Moreover, I think that this retry logic would be an optimization of the existing logic and, thus, not strictly release critical. Since the batch request is anyway only a band aid/temporary bridge, I would like to avoid to add even more logic which we need to rework later on just before the feature freeze.
   
   Please object if you think that this feature is a release blocker and needs to be added.

----------------------------------------------------------------
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:
users@infra.apache.org


With regards,
Apache Git Services