kubernetesschedulingPractitioner

Unlocking Mutable PersistentVolume Node Affinity in Kubernetes v1.35

5 min read Kubernetes BlogJan 8, 2026Reviewed for accuracy

Practitioner — Hands-on experience recommended

Kubernetes v1.35 brings a significant change with mutable PersistentVolume node affinity, addressing a common pain point in resource management. Previously, once set, the node affinity for a PV was immutable, which could lead to inefficiencies when the underlying infrastructure changed. Now, you can modify the node affinity to reflect real-time conditions, enhancing the way your applications interact with the underlying storage.

To leverage this feature, you need to enable the MutablePVNodeAffinity feature gate on the APIServer, which is disabled by default. Once enabled, you can change the PV node affinity using the VolumeAttributesClass API. For example, you might adjust the node affinity from a specific zone to a broader region, allowing for more flexible scheduling. Here's how you might define the node affinity in your PV specification:

YAML

1spec:
2  nodeAffinity:
3    required:
4      nodeSelectorTerms:
5      - matchExpressions:
6        - key: topology.kubernetes.io/zone
7          operator: In
8          values:
9          - us-east1-b

However, there are important considerations to keep in mind. Changing the PV node affinity does not alter the actual accessibility of the underlying volume. There's also a race condition when tightening node affinity; the scheduler might still place a Pod on an old node that can no longer access the volume. If you update the PV and immediately start new Pods in a script, you may encounter unexpected behavior. Always ensure the underlying volume is updated in the storage provider before making changes to the PV affinity.

This feature is currently in alpha, so be cautious about its stability and potential edge cases in production environments.

Key takeaways

→Enable the MutablePVNodeAffinity feature gate on the APIServer to use this functionality.
→Adjust PV node affinity dynamically to optimize resource allocation and scheduling.
→Remember that changing PV node affinity does not change the volume's accessibility.
→Be aware of race conditions when tightening node affinity; the scheduler may not reflect changes immediately.
→Update the underlying volume in the storage provider before modifying PV affinity.

Why it matters

This feature allows for more dynamic and efficient resource management in Kubernetes, which can lead to better application performance and reduced downtime during infrastructure changes.

Code examples

YAML

1spec:
2  nodeAffinity:
3    required:
4      nodeSelectorTerms:
5      - matchExpressions:
6        - key: topology.kubernetes.io/zone
7          operator: In
8          values:
9          - us-east1-b

YAML

1spec:
2  nodeAffinity:
3    required:
4      nodeSelectorTerms:
5      - matchExpressions:
6        - key: topology.kubernetes.io/region
7          operator: In
8          values:
9          - us-east1

YAML

1spec:
2  nodeAffinity:
3    required:
4      nodeSelectorTerms:
5      - matchExpressions:
6        - key: provider.com/disktype.gen1
7          operator: In
8          values:
9          - available