Github user sryza commented on the pull request:
https://github.com/apache/spark/pull/2746#issuecomment-58722547
This is looking good @andrewor14 . Posted a few comments inline and had a
broader question:
Would you mind motivating the Timer machinery a little bit? It took me a
while to understand how to reason about, and I'm wondering whether a simpler
approach would make sense. IIUC, it also means that every time we submit a
task against an idle cluster, even if it gets dispatched quickly, we end up
scheduling and then cancelling timer tasks. Would it make sense to have
noMorePendingTasks and newPendingTasks simply update a variable that holds the
last time we transitioned from no pending tasks to pending tasks? And then to
have a single timer that checks this variable periodically and makes decisions
based on it?
---
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
