Skip to content

YARN-11732. Fix potential NPE when calling SchedulerNode#reservedContainer for CapacityScheduler #7065

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

Merged
Hexiaoqiao merged 1 commit into from
Oct 16, 2024
Merged

YARN-11732. Fix potential NPE when calling SchedulerNode#reservedContainer for CapacityScheduler #7065

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
3 changes: 3 additions & 0 deletions ...oop/yarn/server/resourcemanager/monitor/capacity/ReservedContainerCandidatesSelector.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -170,6 +170,9 @@ private NodeForPreemption getPreemptionCandidatesOnNode(
Map<ApplicationAttemptId, Set<RMContainer>> selectedCandidates,
Resource totalPreemptionAllowed, boolean readOnly) {
RMContainer reservedContainer = node.getReservedContainer();
if (reservedContainer == null) {
return null;
}
Resource available = Resources.clone(node.getUnallocatedResource());
Resource totalSelected = Resources.createResource(0);
List<RMContainer> sortedRunningContainers =
Expand Down
5 changes: 2 additions & 3 deletions ...n/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/AbstractYarnScheduler.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -757,10 +757,9 @@ private void completeOustandingUpdatesWhichAreReserved(
RMContainer rmContainer, ContainerStatus containerStatus,
RMContainerEventType event) {
N schedulerNode = getSchedulerNode(rmContainer.getNodeId());
if (schedulerNode != null &&
schedulerNode.getReservedContainer() != null) {
if (schedulerNode != null) {
RMContainer resContainer = schedulerNode.getReservedContainer();
if (resContainer.getReservedSchedulerKey() != null) {
if (resContainer != null && resContainer.getReservedSchedulerKey() != null) {

@zeekling zeekling Sep 28, 2024

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

It is recommended to add a try catch module. I have made this change in the production environment before, but it still reports a null pointer.

@TaoYang526 TaoYang526 Sep 29, 2024 •
edited
Loading

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

Thanks @zeekling for the review.
I'm not sure why your environment still reported NPE after changing like that, resContainer.getReservedSchedulerKey() won't throw NPE any more with the previous not-null check: if resContainer is null. Could you please attach some details of the NPE and your changes?
For this change, I think NPE should be fixed instead of be caught in general.

@zeekling zeekling Oct 2, 2024

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

I modified like your change.

@TaoYang526 TaoYang526 Oct 8, 2024

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

I can't find any chance to reproduce the NPE, could you please explain in which case NPE can be thrown again? If we can locate that risk point, I would like to fix it instead of catching it in try block. Thanks.

ContainerId containerToUpdate = resContainer
.getReservedSchedulerKey().getContainerToUpdate();
if (containerToUpdate != null &&
Expand Down
5 changes: 3 additions & 2 deletions ...p/yarn/server/resourcemanager/scheduler/capacity/allocator/RegularContainerAllocator.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -858,12 +858,13 @@ private ContainerAllocation allocate(Resource clusterResource,
FiCaSchedulerNode node = iter.next();

// Do not schedule if there are any reservations to fulfill on the node
RMContainer nodeReservedContainer = node.getReservedContainer();
if (iter.hasNext() &&
node.getReservedContainer() != null &&
nodeReservedContainer != null &&
isSkipAllocateOnNodesWithReservedContainer()) {
LOG.debug("Skipping scheduling on node {} since it has already been"
+ " reserved by {}", node.getNodeID(),
node.getReservedContainer().getContainerId());
nodeReservedContainer.getContainerId());
ActivitiesLogger.APP.recordSkippedAppActivityWithoutAllocation(
activitiesManager, node, application, schedulerKey,
ActivityDiagnosticConstant.NODE_HAS_BEEN_RESERVED, ActivityLevel.NODE);
Expand Down
8 changes: 4 additions & 4 deletions ...org/apache/hadoop/yarn/server/resourcemanager/scheduler/common/fica/FiCaSchedulerApp.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -520,13 +520,13 @@ public boolean accept(Resource cluster,
// When reserve a resource (state == NEW is for new container,
// state == RUNNING is for increase container).
// Just check if the node is not already reserved by someone
if (schedulerContainer.getSchedulerNode().getReservedContainer()
!= null) {
RMContainer reservedContainer =
schedulerContainer.getSchedulerNode().getReservedContainer();
if (reservedContainer != null) {
if (LOG.isDebugEnabled()) {
LOG.debug("Try to reserve a container, but the node is "
+ "already reserved by another container="
+ schedulerContainer.getSchedulerNode()
.getReservedContainer().getContainerId());
+ reservedContainer.getContainerId());
}
return false;
}
Expand Down