Live
Agentic Development Shifts Engineer Responsibilities: Practical ImplicationsManaging GPU Capacity with AWS Batch Quota Management for SageMaker TrainingFrom Demos to Code‑Free Labs: How the UK Amazon Quick User Group’s Shift Impacts EngineersHardware Audit Logging for Azure Cloud HSM: Practical Implications for EngineersAzure Foundry Copilot Harness: Managed Agent Execution for EngineersPractical Takeaways from KubeCon End‑User Sessions for Cloud‑Native EngineersBinding Cloudflare Workers to Specific VPCs per ConfigurationAgentic Regression Testing with Cached Playwright Replays: Implications for CI PipelinesAgentic Development Shifts Engineer Responsibilities: Practical ImplicationsManaging GPU Capacity with AWS Batch Quota Management for SageMaker TrainingFrom Demos to Code‑Free Labs: How the UK Amazon Quick User Group’s Shift Impacts EngineersHardware Audit Logging for Azure Cloud HSM: Practical Implications for EngineersAzure Foundry Copilot Harness: Managed Agent Execution for EngineersPractical Takeaways from KubeCon End‑User Sessions for Cloud‑Native EngineersBinding Cloudflare Workers to Specific VPCs per ConfigurationAgentic Regression Testing with Cached Playwright Replays: Implications for CI Pipelines
Kubernetes

Changed Block Tracking API Hits Beta – Migration Steps for Kubernetes Operators

AI SummaryPowered by AI

The SnapshotMetadataService CRD for Changed Block Tracking has moved from v1alpha1 to v1beta1, removing the Alpha version and requiring a manual manifest and code update. This change impacts driver operators and backup tool developers who must align with the new API to continue using CBT in Kubernetes 1.33+ environments.

The snapshot metadata service for Changed Block Tracking (CBT) has been promoted from Alpha to Beta with the March 2026 v1.0.0 release of the external‑snapshot‑metadata project. The promotion replaces the v1alpha1 SnapshotMetadataService CRD with a v1beta1 version, removes the old API, and requires a manual update of manifests and client code.

Version bump and migration steps

All drivers and controllers that interact with the CBT API must now reference the CRD at cbt.storage.k8s.io/v1beta1. The schema itself has not changed, but the previous v1alpha1 definition is no longer served. Practitioners need to:

  • Apply the CRD manifest bundled with the v1.0.0 release.
  • Change any SnapshotMetadataService resources to use apiVersion: cbt.storage.k8s.io/v1beta1.
  • Update any code that watches or mutates the CRD to target the new version.

There is no automatic conversion; the change is a one‑time manual operation.

Compatibility and required components

Beta support is gated by a minimum Kubernetes version of 1.33 and a CSI spec version of 1.10 or newer. The sidecar image to run alongside a CSI driver is registry.k8s.io/sig-storage/csi-snapshot-metadata:v1.0.0. Operators must ensure their drivers ship this sidecar and expose the SnapshotMetadata gRPC service.

Operational considerations

CBT continues to apply only to block‑level volumes; file‑based volumes are out of scope. When deploying a driver, the workflow remains the same as in Alpha: install the CRD, create a SnapshotMetadataService resource, and invoke the GetMetadataAllocated and GetMetadataDelta RPCs via a client such as snapshot-metadata-lister. The host‑path driver example still demonstrates a full end‑to‑end flow.

From an operations perspective, the removal of the Alpha API eliminates the need to run two parallel CRD versions, simplifying version management. However, the manual update introduces a brief window where mismatched API versions could cause controller errors. Teams should schedule the migration during a maintenance window and verify that all custom controllers and backup tools have been recompiled or re‑configured against the new API.

Related CloudNinjas coverage: Kubernetes.

What This Means For Practitioners

For CSI driver maintainers, the beta status signals that the API is stable enough for broader adoption, making it a good time to add CBT support if it aligns with your storage strategy. Backup and data‑protection developers should start testing the streaming client and iterator package against the v1beta1 API and provide feedback before the feature reaches GA. Finally, monitor the SIG Storage and Data Protection Working Group discussions for upcoming compatibility or security guidance that could affect deployment pipelines.

Originally published atKubernetes Blog