PrivateAIM Hub

Appearance

Kubernetes (Helm) ​

FLAME Hub can be deployed to Kubernetes using the official Helm chart. This is the recommended approach for production environments.

TIP

Deployment manifests, scripts, and step-by-step guides are in the hub-deployment repository under kubernetes/.

Prerequisites ​

  • A functioning Kubernetes cluster with ingress capability
  • A default storage class configured
  • Helm installed
  • Two domain names: one for Hub, one for Harbor (can use /etc/hosts for local setups)

Quick Start ​

bash
# Add the Helm repository
helm repo add flame https://PrivateAIM.github.io/helm
helm repo update

# Install with custom values
helm install hub -f values.yaml flame/hub

Kubernetes Distributions ​

bash
# Automated setup
git clone https://github.com/PrivateAIM/hub-deployment.git
cd hub-deployment
bash scripts/1_microk8s_setup.sh

The setup script installs MicroK8s and enables the required addons:

  • dashboard — Kubernetes dashboard
  • ingress — Nginx ingress controller
  • hostpath-storage — local persistent volumes
  • metrics-server — resource monitoring

Default storage path: /var/snap/microk8s/common

Minikube (Development/Testing) ​

bash
minikube start
minikube addons enable ingress

For DNS resolution, configure your local hosts file or use the provided scripts/minikube-dns.ps1 (Windows).

Configuration ​

Create a custom values file rather than modifying defaults. Reference the upstream values.yaml for all available options.

Example override (values.yaml):

yaml
global:
  flameHub:
    ingress:
      enabled: true
      ssl: true
      hostname: "hub.local"

harbor:
  enabled: true
  externalURL: "https://harbor.hub.local/"

Ingress Options ​

Path-based routing (default): All Hub services are routed via path prefixes under a single hostname. Harbor requires a separate hostname.

Gateway API with F5 NGINX Gateway Fabric: Recommended for setups that involve large file uploads (better chunked transfer support).

Storage Replication (Optional) ​

For multi-node clusters (3+ nodes), OpenEBS Mayastor provides replicated persistent storage with automatic failover.

Node Preparation ​

bash
# Automated setup (run on each node)
bash scripts/2_prepare_for_mayastor.sh

This script:

  1. Enables 1024 hugepages (2MB each)
  2. Loads the nvme-tcp kernel module
  3. Labels the node with openebs.io/engine=mayastor

Installation ​

Follow the OpenEBS Mayastor guide in the hub-deployment repository to:

  1. Clone the FLAME Helm repository
  2. Install the OpenEBS wrapper chart
  3. Configure the mayastor-replicated storage class in your Hub values

Further Reading ​