Github user squito commented on the issue:
https://github.com/apache/spark/pull/15505
Looks like @kayousterhout posted some comments addressing my concerns on
https://github.com/apache/spark/pull/16053 at the same time as my last set of
comments. But essentially it sounds like Kay has done the performance
measurements to verify her approach works now, and it looks cleaner to me.
Kay -- earlier you had said:
> I'd propose first merging #16053 (I'll clean this up to include the Mesos
changes) and then merging this PR
But if I understand correctly, you are really recommending closing this,
and going with https://github.com/apache/spark/pull/16053 plus your
still-to-be-shared solution to
https://issues.apache.org/jira/browse/SPARK-18890?
@witgo do you have any concerns with that approach? Your work in
discovering the issue and suggesting a solution are definitely appreciated, but
I think Kay's proposal is a little cleaner. (its also fine if you think there
is a a benefit to your proposal and would like to keep pushing it, I'd just
like to set expectations.)
In any case, I think I will wait on doing any further review until Kay
finishes the issues she mentioned for
https://github.com/apache/spark/pull/16053 *and* shares a proposal for
https://issues.apache.org/jira/browse/SPARK-18890 since I think they may need
to be evaluated together to some extent, even if they end up getting merged
independently -- lemme knwo if you think that plan doesn't make sense for any
reason.
---
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]
