Skip to content

[SPARK-41410][K8S][FOLLOWUP] Skip splitSlots and requestNewExecutors in case of 0 snapshot #39079

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
dongjoon-hyun wants to merge 2 commits into from
Closed

[SPARK-41410][K8S][FOLLOWUP] Skip splitSlots and requestNewExecutors in case of 0 snapshot #39079

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
6ad0da8
[SPARK-41410][K8S][FOLLOWUP] Skip splitSlots and requestNewExecutors …
dongjoon-hyun Dec 15, 2022
969eaeb
Address comment
dongjoon-hyun Dec 15, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 6 additions & 1 deletion ...es/core/src/main/scala/org/apache/spark/scheduler/cluster/k8s/ExecutorPodsAllocator.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -152,6 +152,7 @@ class ExecutorPodsAllocator(
applicationId: String,
schedulerBackend: KubernetesClusterSchedulerBackend,
snapshots: Seq[ExecutorPodsSnapshot]): Unit = {
logDebug(s"Received ${snapshots.size} snapshots")
Copy link
Member Author

@dongjoon-hyun dongjoon-hyun Dec 15, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I added this as DEBUG level because Received 0 snapshots is too verbose.

val k8sKnownExecIds = snapshots.flatMap(_.executorPods.keys).distinct
newlyCreatedExecutors --= k8sKnownExecIds
schedulerKnownNewlyCreatedExecs --= k8sKnownExecIds
Expand Down Expand Up @@ -352,8 +353,12 @@ class ExecutorPodsAllocator(
}
}

// Try to request new executors only when there exist remaining slots within the maximum
// number of pending pods and new snapshot arrives in case of waiting for releasing of the
// existing PVCs
val remainingSlotFromPendingPods = maxPendingPods - totalNotRunningPodCount
if (remainingSlotFromPendingPods > 0 && podsToAllocateWithRpId.size > 0) {
if (remainingSlotFromPendingPods > 0 && podsToAllocateWithRpId.size > 0 &&
!(snapshots.isEmpty && podAllocOnPVC && maxPVCs <= PVC_COUNTER.get())) {
Comment on lines 359 to +361
Copy link
Member

@viirya viirya Dec 15, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This condition is getting complicated now. Can we add a comment on it?

Copy link
Member Author

@dongjoon-hyun dongjoon-hyun Dec 15, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sure, @viirya

ExecutorPodsAllocator.splitSlots(podsToAllocateWithRpId, remainingSlotFromPendingPods)
.foreach { case ((rpId, podCountForRpId, targetNum), sharedSlotFromPendingPods) =>
val numMissingPodsForRpId = targetNum - podCountForRpId
Expand Down