# Volcano v1.12.2

2025-08-14 · [Original release notes](https://github.com/volcano-sh/volcano/releases/tag/v1.12.2) · [Full summary](https://ratatosk.io/en/releases/volcano/v1.12.2)

Breaking changes 2 · Important fixes 4 · Enhancements 1 · Dependency updates 1 · Other changes 3

## Breaking changes (2)

### HyperNode `NodeCount` now counts distinct matching Nodes

- Affects: installations reconciling HyperNode resources

Previously, the count included every `HyperNode.Spec.Members` entry, even if it did not match an existing Node. If listing Nodes fails, the controller logs the error and reports zero.

Evidence: “+ nodes, err := hn.nodeLister.List(labels.Everything())” https://github.com/volcano-sh/volcano/pull/4488

> * Automated cherry pick of #4473: fix node count reconcile by @Monokaix in https://github.com/volcano-sh/volcano/pull/4488

### Hypernode controller release-name key changed to `HELM_RELEASE_NAME`

- Affects: hypernode controller configurations using the release-name environment variable (Inferred)

Previously, `ReleaseNameEnvKey` was defined as `RELEASE_NAME`.

Evidence: “- ReleaseNameEnvKey = "RELEASE_NAME"” https://github.com/volcano-sh/volcano/pull/4490

> * [cherry-pick for 1.12]Fix incorrect definition of ReleaseNameEnvKey by @ouyangshengjia in https://github.com/volcano-sh/volcano/pull/4490

## Important fixes (4)

### Negative-score selection fixed in `SelectBestNodeAndScore` and `SelectBestHyperNode`

- Affects: node and hypernode plugin scoring (Inferred)

The selectors can now choose the highest-scoring available candidate even when scores are below `-1`.

> * [cherry-pick for 1.12]Fix the issue where SelectBestNode returns nil when plugin scores are negative by @guoqinwill in https://github.com/volcano-sh/volcano/pull/4472

### Hierarchical queue capacity metrics now updated

- Affects: capacity plugin setups with hierarchical queues and queue-metric monitoring (Inferred)

When the capacity plugin built attributes for hierarchical queues, their capacity and real-capacity metrics were missing. Processing now updates those metric series with each queue's capability and real-capability resources.

> * Automated cherry pick of #4487: Add missing capacity metrics in hierarchical queues by @JesseStutler in https://github.com/volcano-sh/volcano/pull/4494

### `StatefulSet` `PodGroup` synchronization fixed

- Affects: StatefulSets with replicas greater than zero and matching pods scheduled by a configured Volcano scheduler (Inferred)

When a StatefulSet is added or updated and its pod template changes, the controller creates a missing `PodGroup` or updates an existing one’s spec, labels, and annotations to match the pod.

> * [Cherry-pick] fix: update podGroup when statefulSet update by @Poor12 in https://github.com/volcano-sh/volcano/pull/4522

### Kube-scheduler `Goroutines` metric initialization panic fixed

- Affects: Volcano scheduler setups with `EnableMetrics` or `EnablePprof` enabled

Calling kube-scheduler-related plugins could panic because the `Goroutines` metric was not initialized in `server.go`. The scheduler now initializes the metric in the server startup path before starting the metrics server.

> * Automated cherry pick of #4422: Move kube-scheduler related metrics initilization to server.go to avoid panic by @JesseStutler in https://github.com/volcano-sh/volcano/pull/4461

## Enhancements (1)

- README monitoring installation switches to the installer manifest

## Dependency updates (1)

1 more, none mentioning security

- `* Automated: Bump version to v1.12.2 by @JesseStutler in https://github.com/volcano-sh/volcano/pull/4518`

## Other changes (3)

- Release workflows and build defaults use version configuration
- Job controller deep-copies `PodGroup`s for updates
- Volcano Helm chart and development image references updated to v1.12.2

---

Summaries are produced by AI, so they can be wrong, and they are provided without warranty.
