Github user mateiz commented on the pull request:
https://github.com/apache/spark/pull/3816#issuecomment-68793027
@sryza any progress investigating this? FYI those NO_PREF tasks can be
returned when the locality level is "any" and "rack_local" too. See
TaskLocality.scala. If you don't show a concrete benefit from this, I'd suggest
closing the issue and not changing it. The way I see it, this could break the
scheduling order in existing jobs, with minimal benefit in performance (just
avoiding one small delay for datasets that include both combined and
non-combined files).
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
