Skip to content

Kubernetes Storage

Simplyblock runs natively on Kubernetes, where it is deployed and managed by the Simplyblock Operator through Custom Resource Definitions (CRDs) of the API group storage.simplyblock.io/v1alpha2. Workloads consume storage through the simplyblock CSI driver using standard Kubernetes objects such as StorageClass, PersistentVolumeClaim, and VolumeSnapshot. Provisioning, snapshotting, cloning, and resizing therefore become part of the regular application lifecycle instead of a separate administrative task.

The installation is driven by custom resources. A single Helm chart installs the operator together with a ControlPlane and a SimplyblockDriver resource, and the operator installs the control plane and the CSI driver from them. The operator then discovers the workers and their devices and writes a draft ClusterDeploymentConfig. Once an administrator reviews and approves that draft, the operator creates the storage cluster, its storage nodes, a default storage pool, and a matching StorageClass.

Entities such as StorageCluster, StorageNode, and StoragePool describe the desired state. Imperative operations, for example, a restart, a rolling restart, a node migration, or a restore, are requested with dedicated *Ops resources (StorageClusterOps, StorageNodeOps, ControlPlaneOps, and others) that report their progress in their own status.

Both hyper-converged and disaggregated topologies are supported. In a hyper-converged deployment, simplyblock storage services share Kubernetes worker nodes with application workloads. In a disaggregated deployment, they run on dedicated workers, by default within the same cluster, or optionally in a separate storage cluster. A wide range of Kubernetes distributions is supported, including OpenShift, SUSE Rancher (RKE2 and K3s), and Talos.

Before deploying, review the node sizing, network, and erasure coding guidance in Deployment Preparation. That guidance applies to all deployment models, with additional sizing notes for hyper-converged clusters where compute and storage share nodes.

Coming soon

A centralized hub deployment is in development. In this model, the Simplyblock Operator and the control plane run on a hub cluster that manages the storage of many member Kubernetes clusters through Open Cluster Management (OCM). See Control Plane and Operator.

What This Section Covers

  • Install: Install the operator with its Helm chart, let it install the control plane and the CSI driver, create a storage cluster from a discovered and approved deployment configuration, and apply distribution-specific configuration.
  • Usage: Configure storage classes and quality of service, and provision, expand, snapshot, clone, encrypt, and remove logical volumes using Kubernetes resources.
  • Operations: Monitor cluster health, scale capacity, protect volumes with backups, replicate asynchronously, coordinate node drains, and run cluster operations through the operator.