Skip to content

Virtualization Overview

Virtualization is a core part of OnyxNet, providing environments for self-hosted applications, network services, infrastructure monitoring, and experimentation.

The infrastructure is organized around two Proxmox environments with distinct responsibilities:

  • Aesir: Three-node production cluster
  • Vanir: Single-node experimental environment

Both environments support hands-on infrastructure engineering while serving different operational goals.

Attribute Aesir Vanir
Primary role Production services Testing and experimentation
Platform Proxmox VE Proxmox VE
Physical nodes 3 1
Hardware HP EliteDesk 800 G5 Mini Intel NUC5i5MYHE
Cluster configuration Three-node cluster Single-node environment
Workload priority Operational stability Flexibility and experimentation

Aesir is the primary virtualization platform for services intended to operate reliably in the homelab.

Production workloads should be managed with attention to resource consumption, backups, dependencies, and recovery requirements.

The cluster contains three similarly configured physical systems.

Explore Aesir

Vanir provides a separate environment for testing applications, exploring new configurations, and evaluating changes before considering them for production use.

Its single-node design keeps experimentation separate from the primary production cluster.

Explore Vanir

Proxmox VE provides the virtualization management platform for both environments.

Workloads can use:

  • Virtual machines: Full guest operating systems
  • Linux containers: Lightweight Linux workloads

The appropriate deployment method depends on the application’s isolation, operating system, resource, and compatibility requirements.

ZFS is part of the storage architecture used by the Aesir environment.

An important architectural distinction is that similarly named ZFS pools on different nodes do not automatically constitute shared storage.

Storage locality affects migration, replication, availability, and recovery planning.

Proxmox Backup Server is part of the existing homelab backup infrastructure.

Backup coverage, retention policies, and restoration procedures should be validated before presenting a workload as fully protected.

Separating workloads by operational purpose reduces the need to perform experimental changes inside the primary production environment.

Each workload should have resource allocations appropriate to its actual requirements.

Monitoring and performance testing should inform CPU, memory, and storage decisions.

A multi-node cluster does not automatically guarantee application high availability.

Availability depends on workload configuration, storage accessibility, cluster behavior, failure handling, and tested recovery procedures.

Infrastructure modifications should identify:

  1. The problem or requirement.
  2. The proposed change.
  3. Expected dependencies and risks.
  4. A validation method.
  5. A rollback or recovery approach.
  6. The final outcome.

The existing environments are being evaluated for improved workload placement, resource utilization, storage management, and reliability.

The development objective is to reuse existing hardware rather than rebuild the environments or require immediate hardware purchases.

This documentation will distinguish verified operational configurations from proposed changes.