Operating System
While choosing an operating system for a homelab is not a purely technical decision, is also a statement about how you want to work, what you trust, and how much friction you are willing to live with. For February, that choice is Proxmox VE, and it is not a first date.
Proxmox and I have history. This is not a case of picking something unfamiliar and hoping for the best. It is a deliberate return, going in with a clearer sense of what works, what does not, and what to do differently this time.
Why Proxmox
The honest answer is that Proxmox sits in a sweet spot that is genuinely hard to find elsewhere. It is a Type 1 hypervisor, which means it runs directly on the hardware rather than on top of another OS, and it brings both KVM virtual machines and LXC containers under one roof with a web interface that is functional without being patronising.
It is also open source and free to use without a subscription, though the enterprise repository is locked behind one. For a homelab, the community repository is fine. That is a trade-off worth making.
There are alternatives. TrueNAS makes sense if storage is the primary concern. Unraid has a passionate following. Docker on a bare OS is always an option for those who prefer simplicity. But Proxmox offers something those approaches do not always: genuine flexibility without asking you to fight the tool to get it.
The Approach: LXC First
The working principle for February is LXC containers where possible, virtual machines where necessary. Containers are leaner, faster to spin up, and share the host kernel rather than emulating their own hardware stack. For most services, that is exactly what you want.
Virtual machines earn their place when isolation matters more than efficiency, when something needs its own kernel, or when a workload is awkward to containerise cleanly. The distinction is worth keeping deliberate rather than defaulting to one or the other out of habit.
Planned Services
February is starting focused rather than sprawling. The initial workload is built around a small, purposeful set of services.
Nextcloud is the anchor. Bringing files, calendars, contacts, and more back under direct control is one of the core reasons February exists. It will get its own container, its own resources, and the attention it deserves as the primary service.
The *arr stack follows close behind, alongside Jellyfin for media serving. Eight 4TB drives need something worth filling them with and a decent way to serve it back out. This combination handles both sides of that neatly.
The expectation is that this list grows over time. That is the nature of a homelab. But starting with a clear, manageable set of services means February gets properly bedded in before the scope creeps, as it inevitably will.