Github user squito commented on the pull request:
https://github.com/apache/spark/pull/6291#issuecomment-137142472
Hi @markhamstra this has become important for a customer. I can pick up
trying to add the right unit test if that would be helpful (it would be a good
exercise for me in any case). I think I can get to it in the next couple of
days.
Can I also ask for some confirmation about the nature of the previous fix?
Would it be correct to say: The previous fix was a strict improvement, and
does eliminate the NPE. However, it means that if you use the FAIR Scheduler,
then job priorities will not be properly respected. If you use the FIFO
scheduler, there is no issue. And even with the FAIR scheduler, it does not
effect correctness at all, just prioritization.
That is not to diminish the important of getting a better fix going
forward, but I'd like to clearly explain what the issue is in previous spark
versions.
---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
