Github user cmccabe commented on the pull request:
https://github.com/apache/spark/pull/1486#issuecomment-50089712
So I changed it to no longer rely on order, and instead rely on an explicit
string that gets added to the returned Seq[String]. I also implemented this
for NewHadoopRDD.
One issue that this change still has is that it isn't integrated into how
we launch executors. We'd like to launch executors on hosts where the input
RDD is cached; so far, I haven't implemented this.
---
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.
---
