Aviator: VM Provisioning Kubernetes Controller

A custom Kubernetes controller for automating OS provisioning in cloud environments, with integrated support for the Naver Cloud Platform (NCP) through the Aviator service SDK. The Aviator-service repository contains the core SDK implementation for cloud infrastructure automation.

Overview

This Kubernetes controller automates the provisioning and management of virtual machines through custom resources. It integrates with NCP APIs to handle VM lifecycle operations including creation, updates, deletion, and status monitoring.

Features

• Custom Resource Definitions (CRDs)

  • Group: vm.cloudclub.io
  • Version: v1
  • Resources:
    • Provision: Manages VM provisioning lifecycle
    • Plan: Handles VM planning and configuration

• Integrated NCP Operations

  • Server instance creation and management
  • Network interface configuration
  • Block storage management
  • Access control group handling
  • Status monitoring and updates

Architecture

Component Structure

1. Controller Core

   • Custom Resource Definitions
   • Reconciliation Logic
   • Status Management

2. Aviator Service SDK Integration

   • ServerService: Handles VM operations
   • KeyService: Manages authentication
   • Error handling and status checking
   • NCP API integration

Server Operations Interface

type ServerInterface interface {
    List(url string, request *types.ListServerRequest) (*types.ListServerResponse, error)
    Create(url string, request *types.CreateServerRequest, params []int) (*types.CreateServerResponse, error)
    Update(url string, request *types.UpdateServerRequest) (*types.UpdateServerResponse, error)
    Start(url string, request *types.StartServerRequest) (*types.StartServerResponse, error)
    Stop(url string, request *types.StopServerRequest) (*types.StopServerResponse, error)
    Delete(url string, request *types.DeleteServerRequest) (*types.DeleteServerResponse, error)
}

Custom Resources

Provision Resource Example

apiVersion: vm.cloudclub.io/v1
kind: Provision
spec:
  regionCode: KR
  server:
    imageProductCode: UBUNTU20
    productCode: SPSVRSTAND000056
  vpcNo: "vpc-123"
  subnetNo: "subnet-456"
  phase: Create

Development

Prerequisites

• Go 1.16+
• Kubernetes cluster (1.19+)
• Docker or compatible container runtime
• NCP Account and API Credentials
• Aviator Service SDK

Setting Up Development Environment

1. Clone the repositories:

# Main controller
git clone https://github.com/your-org/vm-controller
# Aviator Service SDK
git clone https://github.com/cloud-club/Aviator-service

2. Configure NCP credentials:

export NCP_ACCESS_KEY="your-access-key"
export NCP_SECRET_KEY="your-secret-key"

Build and Test

# Build the controller
make build

# Run tests
make test

# Run linter
make lint

CI/CD Pipeline

The project uses GitHub Actions for CI/CD with the following stages:

1. Build & Test

   • Code compilation
   • Unit tests
   • Integration tests
   • Code coverage reporting

2. Quality Gates

   • Linting
   • Security scanning
   • Dependency checking

3. Deployment

   • Container image building
   • Image scanning
   • Registry push
   • Kubernetes deployment

Pipeline Configuration

# Example workflow stages
name: CI/CD Pipeline

on:
  push:
    branches: [ main ]
  pull_request:
    branches: [ main ]

jobs:
  build:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v2
      - name: Build
        run: make build
      - name: Test
        run: make test

Usage

1. Installation

# Install CRDs
make install

# Deploy controller
make deploy IMG=<your-registry>/controller:tag

2. Create a VM Instance

apiVersion: vm.cloudclub.io/v1
kind: Provision
metadata:
  name: example-vm
spec:
  phase: Create
  server:
    imageProductCode: "UBUNTU20"
    productCode: "SPSVRSTAND000056"
  vpcNo: "vpc-123"
  subnetNo: "subnet-456"

3. Monitor Status

kubectl get provisions
kubectl describe provision example-vm

Contributing

1. Fork both repositories (controller and Aviator-service)
2. Create feature branches
3. Run tests and linting
4. Submit Pull Requests to both repos if needed
5. Ensure CI pipeline passes