Skip to main content
This guide walks you through installing vCluster using Helm, the recommended method for production deployments.

​
Prerequisites

Before installing vCluster, ensure you have:
  • Kubernetes 1.18+ cluster (host cluster)
  • Helm 3.10.0+ installed locally
  • kubectl configured to access your host cluster
  • Appropriate permissions to create namespaces and cluster-scoped resources

​
Installation Methods

​
Quick Installation

For a basic vCluster installation with default settings:
1

Add the Helm Repository

Add the Loft.sh Helm repository to your Helm client:
helm repo add loft-sh https://charts.loft.sh
helm repo update
2

Install vCluster

Install vCluster in a new namespace:
helm upgrade --install my-vcluster loft-sh/vcluster \
  --namespace vcluster-my-vcluster \
  --create-namespace
This creates a virtual cluster named my-vcluster in the vcluster-my-vcluster namespace.
3

Verify Installation

Check that the vCluster pod is running:
kubectl get pods -n vcluster-my-vcluster
You should see output similar to:
NAME              READY   STATUS    RESTARTS   AGE
my-vcluster-0     1/1     Running   0          2m

​
Custom Installation with Values

For production deployments, create a custom values.yaml file:
1

Create Values File

Create a values.yaml file with your configuration:
values.yaml
# Enable high availability
controlPlane:
  statefulSet:
    highAvailability:
      replicas: 3
    
    # Resource limits for production
    resources:
      limits:
        cpu: 2000m
        memory: 4Gi
      requests:
        cpu: 500m
        memory: 1Gi
    
    # Persistent storage
    persistence:
      volumeClaim:
        enabled: true
        size: 10Gi
        storageClass: "fast-ssd"
  
  # Use external etcd for better performance
  backingStore:
    etcd:
      deploy:
        enabled: true
        statefulSet:
          highAvailability:
            replicas: 3
          resources:
            requests:
              cpu: 200m
              memory: 512Mi
          persistence:
            volumeClaim:
              enabled: true
              size: 10Gi

# Configure resource syncing
sync:
  toHost:
    pods:
      enabled: true
    services:
      enabled: true
    persistentVolumeClaims:
      enabled: true
2

Install with Custom Values

Install vCluster using your custom configuration:
helm upgrade --install my-vcluster loft-sh/vcluster \
  --namespace vcluster-my-vcluster \
  --create-namespace \
  --values values.yaml
3

Verify Configuration

Verify that your custom settings were applied:
# Check the number of replicas
kubectl get statefulset -n vcluster-my-vcluster

# Check the etcd pods
kubectl get pods -n vcluster-my-vcluster -l app=vcluster-etcd

​
Connecting to vCluster

After installation, connect to your vCluster using the vCluster CLI:
1

Install vCluster CLI (if not already installed)

# macOS / Linux
curl -L -o vcluster "https://github.com/loft-sh/vcluster/releases/latest/download/vcluster-$(uname -s)-$(uname -m)"
chmod +x vcluster
sudo mv vcluster /usr/local/bin
2

Connect to vCluster

vcluster connect my-vcluster -n vcluster-my-vcluster
This command:
  • Establishes a port-forward to the vCluster
  • Updates your kubeconfig to include the vCluster context
  • Switches to the vCluster context
3

Verify Connection

Test that you’re connected to the virtual cluster:
kubectl get namespaces
You should see the default namespaces of a fresh Kubernetes cluster.

​
Installation Options

​
Kubernetes Distribution Selection

vCluster supports different Kubernetes distributions. By default, it uses K3s, but you can configure vanilla Kubernetes: 
controlPlane:
  distro:
    k8s:
      enabled: true
      version: "v1.35.0"
      image:
        registry: ghcr.io
        repository: "loft-sh/kubernetes"
        tag: "v1.35.0"

​
Backing Store Options

vCluster supports multiple backing store configurations:

​
Embedded Database (Default)

controlPlane:
  backingStore:
    database:
      embedded:
        enabled: true

​
Embedded etcd

controlPlane:
  backingStore:
    etcd:
      embedded:
        enabled: true

controlPlane:
  backingStore:
    etcd:
      deploy:
        enabled: true
        statefulSet:
          highAvailability:
            replicas: 3

​
External Database

controlPlane:
  backingStore:
    database:
      external:
        enabled: true
        dataSource: "postgres://username:password@hostname:5432/vcluster"

​
Upgrading vCluster

To upgrade an existing vCluster installation: 
# Update the Helm repository
helm repo update

# Upgrade vCluster
helm upgrade my-vcluster loft-sh/vcluster \
  --namespace vcluster-my-vcluster \
  --values values.yaml \
  --reuse-values
Always backup your data before upgrading. If using persistent volumes, ensure they are properly configured with retention policies.

​
Uninstalling vCluster

To completely remove a vCluster: 
helm uninstall my-vcluster --namespace vcluster-my-vcluster
This removes all Kubernetes components associated with the vCluster.
By default, persistent volume claims are retained even after uninstallation. To delete them, you need to manually remove the PVCs or configure the retention policy.

​
Common Installation Pitfalls

​
Insufficient Permissions

Problem: Installation fails with permission errors. Solution: Ensure your service account has cluster-admin privileges or at minimum:
  • Create/manage namespaces
  • Create cluster roles and bindings
  • Create custom resource definitions (if using integrations)

​
Resource Constraints

Problem: vCluster pods are stuck in Pending state. Solution: Check node resources and adjust resource requests: 
controlPlane:
  statefulSet:
    resources:
      requests:
        cpu: 200m    # Reduced from default
        memory: 256Mi

​
Namespace Conflicts

Problem: “namespace already exists” error during installation. Solution: Either:
  • Use a different namespace name
  • Delete the existing namespace (after backing up any important data)
  • Use helm upgrade --install instead of helm install

​
Storage Class Issues

Problem: PVCs remain in Pending state. Solution: Verify your cluster has a default storage class or specify one explicitly: 
controlPlane:
  statefulSet:
    persistence:
      volumeClaim:
        storageClass: "standard"  # or your cluster's storage class

​
Next Steps