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 2020/01/04 03:22:45 UTC

[GitHub] [flink] xintongsong edited a comment on issue #10608: [FLINK-15300][Runtime] Fix sanity check to not fail if shuffle memory fraction is out of min/max range

xintongsong edited a comment on issue #10608: [FLINK-15300][Runtime] Fix sanity check to not fail if shuffle memory fraction is out of min/max range
URL: https://github.com/apache/flink/pull/10608#issuecomment-570752333
 
 
   @tillrohrmann 
   I would consider `RangeFraction` as a data structure that serves `TaskExecutorResourceUtils` only rather than for general purpose. After all, it is a private internal class of `TaskExecutorResourceUtils` and it already assumes the range is represented in `MemorySize`.
   
   `fraction == 1` might be valid for a general purpose range-fraction. But for `TaskExecutorResourceUtils`, none of the task executor resource related fractions (managed, shuffle, jvm-overhead) should ever be 1. Managed memory or shuffle memory should never takes all of the total Flink memory, and JVM overhead should never takes all of the total process memory.
   
   Thus, IMO, it makes sense for `RangeFraction` to not accept `fraction == 1`.

----------------------------------------------------------------
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