Github user andrewor14 commented on the pull request:
https://github.com/apache/spark/pull/2746#issuecomment-59133473
I have updated the code to use polling rather than Java `Timers`. The new
approach uses only 1 extra thread instead of 1 for each executor, and I find
the logic simpler to implement and reason about.
@sryza This still uses the exponential increase policy where each round
stops and wait until the executors are actually granted by the cluster manager
before continuing. I understand this does not address the concerns you raised
on the corresponding JIRA yet, so I am working out the proper semantics there
myself. Other than that, feel free to review the rest of the code.
---
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
