ISSUE DESCRIPTION Standalone elastic RayCluster self-updates raycluster-generation…
- User
ISSUE DESCRIPTION
Standalone elastic RayCluster self-updates raycluster-generation annotation What happened:
A standalone Kueue-managed elastic RayCluster is continuously updated by Kueue.
The
kueue.x-k8s.io/raycluster-generationannotation tracks the RayCluster's ownmetadata.generation, but remains one value behind. For example:metadata.generation: 273390 kueue.x-k8s.io/raycluster-generation: "273389"A later read showed the same pattern:
metadata.generation: 273462 kueue.x-k8s.io/raycluster-generation: "273461"Kueue logs showed repeated RayCluster reconciles followed by RayCluster update webhook calls from the Kueue controller manager.
This is a self-triggered loop in the RayCluster custom annotation path. For standalone RayClusters,
GetCustomAnnotationscallsGetWorkloadslicingRayClusterCustomAnnotations(ctx, c, j.Object(), podSets, j.Name). That helper fetches the same RayCluster and returns the fetched RayCluster generation as an annotation to write back onto the RayCluster. The metadata patch then advances the RayCluster generation, making the annotation stale again.What you expected to happen:
Kueue should not continuously patch a standalone elastic RayCluster when its replica counts have not changed.
The
raycluster-generationannotation is useful for RayJob and RayService, where the top-level object needs to account for an underlying RayCluster generation. A standalone RayCluster should not write its own generation back onto itself.How to reproduce it (as minimally and precisely as possible):
- Install Kueue v0.16.5 with
ElasticJobsViaWorkloadSlices=true. - Enable the RayCluster integration.
- Create a standalone RayCluster with:
metadata: annotations: kueue.x-k8s.io/elastic-job: "true" labels: kueue.x-k8s.io/queue-name: <local-queue>- Inspect the RayCluster generation and annotation repeatedly:
kubectl get raycluster.ray.io <name> -n <namespace> \ -o jsonpath='{.metadata.generation}{"\t"}{.metadata.annotations.kueue\.x-k8s\.io/raycluster-generation}{"\n"}'Anything else we need to know?:
This may be related to #10160, which added generation-based elastic workload names on the release-0.16 branch.
A fix is to omit
kueue.x-k8s.io/raycluster-generationwhen thejobObjectis the same standalone RayCluster asrayClusterName.Environment:
- Kubernetes version (use
kubectl version):- Client Version: v1.33.9
- Server Version: v1.33.10-gke.1115000
- Kueue version (use
git describe --tags --dirty --always):- Deployed controller image:
kueue:v0.16.5 - Helm chart:
kueue-0.16.5
- Deployed controller image:
- Cloud provider or hardware configuration: GKE
- OS (e.g:
cat /etc/os-release): not collected - Kernel (e.g.
uname -a): not collected - Install tools: Helm
- Others:
ElasticJobsViaWorkloadSlices=trueTopologyAwareScheduling=false
ADDITIONAL ADVICE
Since you are given a git repository, you can use git commands to simplify your work. For example, if you made a mistake and want to revert your changes, you can use
git checkout <file>to restore the file to its original state. You can also reset all changes in the repository usinggit reset --hardcommand. Additionally, you can usegit stashandgit stash popto temporarily save your changes and restore them later. Finally, you can see the changes that you've made (which will be submitted when you callsubmit) usinggit statusorgit diffcommands. However you don't need to usegit addorgit commitbefore submitting your changes. In fact,submitonly submits currently unstaged changes because it usesgit diffto get the changes that need to be submitted.CHECKLIST
Before submitting your solution, please go over the following checklist and make sure you've done everything:
- If an error was reported in the issue description, I have successfully reproduced it.
- If an error was reported in the issue description, I have confirmed that my fix resolves the error.
- I have thought about edge cases and made sure my fix handles them as well.
- I have run existing tests in the repository that might have been affected by the change I made and confirmed that they pass. I want you to list every bullet from this checklist and write your reasoning for why you think you did it or didn't need to.
Repository has been uploaded and your shell is currently at the repository root. Time to solve the issue!
(Current directory: /kueue, current file: none) bash-$
- Install Kueue v0.16.5 with
Assistant
Let me start by understanding the issue:
- A standalone elastic RayCluster is being continuously updated by Kueue because it's writing its own generation back onto itself as an annotation (
kueue.x-k8s.io/raycluster-generation). - This creates a self-triggered loop: the annotation update changes the RayCluster's generation, which makes the annotation stale again, causing another reconciliation.
- The fix should omit the
raycluster-generationannotation when the job object is the same standalone RayCluster.
Let me explore the repository structure first to understand the codebase.
- A standalone elastic RayCluster is being continuously updated by Kueue because it's writing its own generation back onto itself as an annotation (
Assistant
Let me find the relevant code for the RayCluster custom annotations:
Assistant
I need to make one command at a time.
Assistant
Let me look at the
common.gofile for the RayCluster integration.Assistant
The open command requires parameters. Let me try just opening the file.
Assistant
Now let me see the callers of this function to understand the context:
Assistant
Let me check the standalone RayCluster controller for a
GetCustomAnnotationsmethod:Assistant
Let me check the raycluster_controller.go:
Assistant
Let me see the rest of the raycluster_controller.go
Assistant
Let me check the reconciler to see how this annotation is applied
Assistant
I need to use the special command format properly.
Assistant
Let me check if the standalone RayCluster implements
JobWithCustomAnnotations:Assistant
Let me search more carefully for where the standalone RayCluster might implement `GetCustomAnnotations
Assistant
Let me check the raycluster webhook:
Assistant
Let me check the git log for recent changes related to this
Assistant
Let me list all files in the raycluster package
Assistant
Let me find the constant definitions for these
Assistant
Let me check if there's a
GetCustomAnnotationsmethod on the RayCluster typeAssistant
Let me look at the webhook's Default function
Assistant
Let me look at the beginning of common.go
Assistant
Let me find where
GetWorkloadNameExtraPartis calledAssistant
Let me check if there's a
GetCustomAnnotationson theRayClustertype, maybe through compositionAssistant
Let me search more broadly
Assistant
Let me check for
GetWorkloadNameExtraParton RayClusterAssistant
Let me look at the reconciler around line 1437 where
GetWorkloadNameExtraPartis calledAssistant
Let me look at
workloadslicing.EnsureWorkloadSlicesAssistant
Let me look at the workloadslicing.go file to understand the flow better.
Assistant
Let me look at the test for
GetWorkloadslicingRayClusterCustomAnnotationsAssistant
Let me see the rest of the test
Assistant
Let me check the
WorkloadSliceEnabledfunction and the interface to understand the fullAssistant
Let me look at this
Assistant
Let me check recent changes to the raycluster package
Assistant
Let me check all methods defined on
RayClustertypeAssistant
Now I understand the issue. Let me look at the function once more and then implement the fix.
The core problem:
GetWorkloadslicingRayClusterCustomAnnotationsfetches a RayCluster and returns its generation as an annotation. WhenjobObjectis the same RayCluster (standalone case), this creates a self-triggered loop because writing the annotation changes the RayCluster's generation.The fix: omit
RayClusterGenerationAnnotationwhen the fetched RayCluster is the same object as `jobAssistant
Now I need to implement the fix. The approach: compare the UID of the fetched RayCluster with the
jobObject's UID. If they match (same object), omit theRayClusterGenerationAnnotationto avoid the self-triggeredAssistant
Let me use the replace command instead
Assistant
Good, the fix is applied. Now let me verify by running the relevant
Assistant
Let me set up the Go environment first
Assistant
The test is failing because the fake client may not assign UIDs, making both empty. Let me check how the fake client handles UIDs
Assistant
Let me look at the RayCluster test utility
Assistant
Let me look at the RayCluster wrapper
Assistant
The fake client doesn't assign UIDs, so both objects have empty UIDs. Let me use a different approach - check if
jobObjectis itself a*rayv1.RayClusterwith the same nameAssistant
Let me navigate to the right location first
Assistant
Let me use a simpler type assertion approach instead of U
Assistant
Now let me run the tests again
Assistant
All tests pass. Now let me run the full test suite for the raycluster package and also the rayjob and rayservice tests
Assistant
Let me run the rayjob and rayservice tests too
Assistant
Let me also check the rayjob tests specifically
Assistant
Let me verify the final state of the changed
Assistant
Let me also verify the build is clean
Assistant
Let me also check if there's any annotation writing in the webhook that could cause this loop
Assistant
Let me run a broader set of tests to make sure nothing is broken
Assistant
All tests pass. Let me verify the diff
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.
Identical first prompt
Qwen/Qwen3.6-27BCompare side by side →
Qwen/Qwen3.6-27BCompare side by side →
Qwen/Qwen3.6-27BCompare side by side →
Qwen/Qwen3.6-27BCompare side by side →
Qwen/Qwen3.6-27BCompare side by side →
- xiaomi/mimo-v2.5-proCompare side by side →
- xiaomi/mimo-v2.5-proCompare side by side →
- xiaomi/mimo-v2.5-proCompare side by side →
- xiaomi/mimo-v2.5-proCompare side by side →
- xiaomi/mimo-v2.5-proCompare side by side →
- gpt-5.6-solCompare side by side →
- gpt-5.6-solCompare side by side →