Github user susanxhuynh commented on a diff in the pull request:
https://github.com/apache/spark/pull/19374#discussion_r143487275
--- Diff:
resource-managers/mesos/src/main/scala/org/apache/spark/scheduler/cluster/mesos/MesosClusterScheduler.scala
---
@@ -276,8 +276,8 @@ private[spark] class MesosClusterScheduler(
private def recoverState(): Unit = {
stateLock.synchronized {
launchedDriversState.fetchAll[MesosClusterSubmissionState]().foreach
{ state =>
- launchedDrivers(state.taskId.getValue) = state
- pendingRecover(state.taskId.getValue) = state.slaveId
+ launchedDrivers(state.driverDescription.submissionId) = state
--- End diff --
Maybe modify the comments up in L.138-150 ^^ to clarify which data
structures are keyed by submission ID vs. task ID:
- Keyed by submission ID: `launchedDrivers`, `queuedDrivers`,
`pendingRetryDrivers`, `finishedDrivers`
- Keyed by task ID: `pendingRecover`
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
