Github user vanzin commented on a diff in the pull request:

    https://github.com/apache/spark/pull/7786#discussion_r40484168
  
    --- Diff: 
core/src/main/scala/org/apache/spark/scheduler/cluster/CoarseGrainedSchedulerBackend.scala
 ---
    @@ -449,6 +458,13 @@ class CoarseGrainedSchedulerBackend(scheduler: 
TaskSchedulerImpl, val rpcEnv: Rp
        */
       protected def doKillExecutors(executorIds: Seq[String]): Boolean = false
     
    +  /**
    +   * Executor ids which will possibly be preempted by cluster manager.
    +   * @return A set of executor IDs
    +   */
    +  protected def preemptedExecutors: Set[String] = synchronized {
    +    preemptedExecutorIDs.toSet
    --- End diff --
    
    I see you modify it in the test, but I think that's the wrong thing to do. 
Better to avoid this transformation or extra allocation or whatever overhead, 
since this method is called pretty often by the scheduling 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 [email protected] or file a JIRA ticket
with INFRA.
---

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to