Installing decoupled

In a decoupled AZ layout, storage and workload run on separate Superphenix clusters. Dedicated storage clusters can serve one or more workload clusters. This model fits performance-sensitive workloads, shared storage across AZs, and production multi-tier designs.

When to choose decoupled

• Performance: dedicated hardware for Ceph and hypervisor tiers without resource contention.
• Shared storage: multiple workload AZs consume the same storage backend.
• Multi-AZ at scale: pairs with a management plane on a dedicated management cluster.

See Deployment topology for the fully decoupled deployment type and support status.

Prerequisites

• Talos clusters for storage and workload tiers (Manual OS installation or Automated OS installation).
• The management plane installed with Kubernetes API access to every cluster.
• Hardware sized per role. See Hardware requirements.

Installation overview

1. Provision a storage cluster and one or more workload clusters.
2. Register each with a Cluster resource (deploymentTopology: Decoupled, with type: Storage or type: Workload).
3. Add a storage class to the workload cluster.
4. Let the operator sync the Superphenix stack on each cluster.

Installing a workload and storage cluster

To install the storage and workload cluster, upgrade the release with the chart clusters: values. The example below adds two new decoupled clusters in the same availability zone: one Storage cluster and one Workload cluster.

clusters:
  spx-us-west01-storage:
    deploymentTopology: Decoupled
    type: Storage
    region: us-west-1
    availabilityZone: west01
    version: v0.1.0
    connection:
      mode: Remote
      url: https://[STORAGE CLUSTER API ENDPOINT]:6443
      # Any of the following option is valid to connect to the cluster
      credentials:
        bearerToken: ""
        caData: ""
        certData: ""
        keyData: ""
        username: ""
        password: ""
        insecure: false
        serverName: ""
    lifecycle:
      pause: false
      manual: false

  spx-us-west01-workload:
    deploymentTopology: Decoupled
    type: Workload
    region: us-west-1
    availabilityZone: west01
    version: v0.1.0
    connection:
      mode: Remote
      url: https://[WORKLOAD CLUSTER API ENDPOINT]:6443
      # Any of the following option is valid to connect to the cluster
      credentials:
        bearerToken: ""
        caData: ""
        certData: ""
        keyData: ""
        username: ""
        password: ""
        insecure: false
        serverName: ""
    lifecycle:
      pause: false
      manual: false

OPERATOR_VERSION="$(curl -fsSL -o /dev/null -w '%{url_effective}' https://github.com/super-phenix/superphenix/releases/latest | sed 's#.*/v##')"

helm upgrade superphenix-operator \
  "oci://ghcr.io/super-phenix/charts/superphenix-operator:${OPERATOR_VERSION}" \
  --namespace superphenix-system \
  -f spx-clusters-values.yaml

You can monitor the status of the Cluster CR to check how far along the installion is.

You'll then need to add a storage class to the workload cluster.

When the deployment is ready, continue with Accessing Superphenix to open the console and monitor applications in ArgoCD.