✎ Key Insights in 30 Seconds
Field note: A field note on the current personal homelab topology: one ThinkCentre host, Proxmox virtualization, three Ubuntu 26 VMs, and the K3s platform running above them.
Problem Context
A homelab becomes more useful when it represents the complete operational path instead of only a Kubernetes manifest. Hoan's current lab starts with one ThinkCentre physical host running an Intel Xeon E-2286M (8 cores / 16 threads) and 32 GB RAM. Proxmox provides the virtualization layer, and three Ubuntu 26 VMs form the K3s environment.
Architecture & Design
The topology is intentionally compact: hardware at the bottom, Proxmox as the host boundary, three Ubuntu 26 guests as K3s nodes, and cluster services such as Traefik, Longhorn, registry, and self-hosted workloads above them. This makes it possible to practice node lifecycle, storage behavior, ingress, image delivery, and recovery without pretending that three VMs on one host provide the same fault domain as three physical machines.
Key Technical Decisions
- 1Use Proxmox to keep the physical host boundary visible while making VM creation, replacement, and resource allocation repeatable.
- 2Use three Ubuntu 26 guests to exercise multi-node K3s scheduling and cluster operations instead of learning only from a single-node setup.
- 3Treat Longhorn, ingress, registry, and CI/CD as operational components that need observation and recovery practice, not just installation steps.
- 4Describe the single-host limitation explicitly: the lab is excellent for systems practice, but the physical host remains a shared failure domain.
Verified repository evidence
infrastructure · direct
Built and operate a Proxmox-based K3s Homelab on one ThinkCentre host with three Ubuntu 26 VMs, Longhorn storage, Traefik ingress, and automated local CI/CD pipelines.
↳ One ThinkCentre host with Intel Xeon E-2286M (8 cores / 16 threads) and 32 GB RAM, running Proxmox with three Ubuntu 26 VMs. The guests run K3s, Longhorn storage, Traefik Ingress Controller, and automated container deployment pipelines.
Self-Hosted Infrastructure Lab ↗
Key Engineering Takeaways
- The most educational infrastructure boundary is often the whole stack: hardware, hypervisor, guest OS, cluster, storage, and delivery.
- Virtualized nodes are useful for practicing distributed-system behavior, as long as their shared physical failure domain is documented.
- A small owned lab creates feedback that managed Kubernetes documentation cannot provide: disks fill, nodes drift, services fail, and recovery has to be performed.