Github user kayousterhout commented on a diff in the pull request:
https://github.com/apache/spark/pull/15249#discussion_r81410592
--- Diff:
core/src/test/scala/org/apache/spark/scheduler/SchedulerIntegrationSuite.scala
---
@@ -325,7 +326,11 @@ private[spark] abstract class MockBackend(
executorIdToExecutor(task.executorId).freeCores +=
taskScheduler.CPUS_PER_TASK
freeCores += taskScheduler.CPUS_PER_TASK
}
- reviveOffers()
+ // optimization (which is used by the actual backends too) -- don't
revive offers on *all*
+ // executors when a task completes, just on the one which completed
+ val exec = executorIdToExecutor(task.executorId)
+ reviveWithOffers(IndexedSeq(WorkerOffer(executorId =
exec.executorId, host = exec.host,
--- End diff --
why do you need to do this optimization for the test? Is the idea to test
exactly what the actual scheduler does? (the extra complexity doesn't seem
worth it unless the code path is pretty different so worth testing)
---
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]
