ADR-0005: KVM/Libvirt Virtualization Platform Choice

Status

Accepted

Context

Qubinode Navigator required a virtualization platform to create and manage virtual machines for development, testing, and production workloads across different environments. The platform needed to support multiple operating systems (RHEL 8/9, Rocky Linux, Fedora), provide good performance on bare-metal and cloud instances, integrate well with Ansible automation, and offer enterprise-grade features for storage, networking, and resource management. The solution needed to be cost-effective, well-supported, and suitable for both single-node and distributed deployments.

Decision

Adopt KVM (Kernel-based Virtual Machine) with libvirt as the primary virtualization platform, enhanced with kcli for simplified VM lifecycle management. KVM provides hardware-accelerated virtualization on Linux systems, libvirt offers a consistent API for VM management, and kcli provides user-friendly tooling for common virtualization tasks.

Consequences

Positive Consequences

  • Excellent performance through hardware-accelerated virtualization
  • Native Linux integration with minimal overhead
  • Strong ecosystem support and enterprise adoption
  • Cost-effective solution with no licensing fees
  • Comprehensive networking and storage options
  • Good integration with Ansible and automation tools
  • Supports live migration and advanced VM features
  • Active development and long-term support from Red Hat/community

Negative Consequences

  • Linux-only solution, limiting cross-platform compatibility
  • Steeper learning curve compared to desktop virtualization solutions
  • Requires hardware virtualization support (Intel VT-x/AMD-V)
  • More complex networking configuration for advanced scenarios
  • Limited GUI management tools compared to commercial alternatives
  • Requires root privileges for many operations

Alternatives Considered

  1. VMware vSphere/ESXi - Rejected due to licensing costs and vendor lock-in
  2. Proxmox VE - Good option but adds additional management layer complexity
  3. Docker/Podman containers - Insufficient for full OS virtualization requirements
  4. OpenStack - Too complex and resource-intensive for the project’s needs
  5. Xen Hypervisor - Less integrated with Red Hat ecosystem
  6. VirtualBox - Not suitable for production/server environments

Evidence Supporting This Decision

  • Extensive KVM host setup automation in setup.sh and qubinode_navigator.sh
  • Libvirt storage configuration and management
  • Integration with kcli for VM lifecycle management
  • Support for bridge networking and storage pools
  • Ansible playbooks for KVM host deployment
  • Hardware compatibility checks and optimization

Implementation Details

KVM Host Setup 

# From setup.sh - KVM host deployment
function deploy_kvmhost() {
    echo "Deploying KVM Host"
    cd "$HOME"/qubinode_navigator
    sudo -E /usr/local/bin/ansible-navigator run ansible-navigator/setup_kvmhost.yml \
        --extra-vars "admin_user=lab-user" --penv GUID \
        --vault-password-file "$HOME"/.vault_password -m stdout
}

Storage Configuration 

# From load-variables.py - Dynamic storage setup
inventory['create_libvirt_storage'] = True
inventory['create_lvm'] = True
inventory['kvm_host_libvirt_extra_disk'] = disks

Networking Setup 

# Network bridge configuration
inventory['configure_bridge'] = configure_bridge
inventory['kvm_host_interface'] = interface
inventory['kvm_host_ip'] = ip

Kcli Integration 

# From setup.sh - Kcli setup for VM management
function setup_kcli_base() {
    echo "Configuring Kcli"
    source ~/.bash_aliases
    qubinode_setup_kcli
    kcli_configure_images
}

Key Features Utilized

  • Hardware Acceleration: Native KVM support for Intel VT-x/AMD-V
  • Storage Management: LVM-based storage pools with libvirt integration
  • Network Management: Bridge networking for VM connectivity
  • Resource Management: CPU, memory, and disk allocation controls
  • Snapshot Support: VM state management and backup capabilities
  • Live Migration: VM mobility between hosts (when configured)

Integration Points

  • Ansible Automation: Playbooks for KVM host setup and VM management
  • Storage Integration: LVM and filesystem-based storage pools
  • Network Integration: Bridge networking with host network interfaces
  • Monitoring: Integration with system monitoring and alerting
  • Backup: VM snapshot and backup automation

Performance Optimizations

  • Hardware virtualization acceleration enabled
  • Optimized storage configurations with LVM
  • Network bridge setup for minimal overhead
  • CPU and memory tuning for virtualization workloads

Security Considerations

  • Isolated VM environments with proper resource limits
  • Network segmentation through bridge configurations
  • Storage encryption support where required
  • Access control through libvirt permissions

Operational Benefits

  • Standardization: Consistent virtualization platform across all environments
  • Automation: Full lifecycle management through Ansible
  • Scalability: Supports both single-node and distributed deployments
  • Cost Efficiency: No licensing costs for virtualization platform
  • Performance: Near-native performance for virtualized workloads
  • Flexibility: Supports various guest operating systems and configurations
  • ADR-0001: Container-First Execution Model with Ansible Navigator
  • ADR-0002: Multi-Cloud Inventory Strategy with Environment-Specific Configurations
  • ADR-0003: Dynamic Configuration Management with Python
  • ADR-0004: Security Architecture with Ansible Vault and AnsibleSafe

Date

2025-01-09

Stakeholders

  • Infrastructure Team
  • DevOps Team
  • System Administrators
  • Development Teams
  • Project Maintainers