Consolidating five services onto one Proxmox PC is efficient — and it quietly makes that one PC the single point of failure for all five. This page explains the risk in plain terms, why a cluster is not the automatic answer, and the proportionate defences that make a single host a sensible choice instead of a gamble.
Every Proxmox success story has the same first act: one old PC, five VMs, everything tidy. The second act is written by a power supply. A single host is a deliberate trade — efficiency now, shared fate always. Taken with open eyes and the right precautions it is a good trade for many people. Taken blindly, it is how a whole office discovers consolidation on a Monday morning.
When five services ran on five old PCs, one dying PC meant one dead service. Consolidate them onto a single Proxmox host and the maths flips: one power supply, one motherboard, one system disk, one network cable now stand behind everything. None of those parts is exotic — they are the most ordinary failure points in computing — but their failure is now a correlated failure: the file server, the accounts VM, the print helper and the test box all go dark in the same instant, and they all wait for the same repair.
Proxmox's famous answer is a cluster with high availability — VMs restart automatically on a surviving node. It is real, and for a home PC project it is mostly the wrong purchase: a proper cluster wants three nodes for reliable quorum (two-node clusters have split-brain headaches that need extra tricks), shared or replicated storage, and identical care three times over. Buying two extra machines so a home lab survives a failure that a backup and an hour of patience also survives is enthusiasm, not engineering. Spend in proportion to what downtime actually costs you.
Home lab and learning: a single host is unambiguously fine. The worst realistic day is restoring from backup and losing some hours of tinkering. Small office, non-critical services — internal tools, a test environment, a file staging box: fine too, provided the backups-off-box and restore-order rules above are actually followed. The one system the business cannot operate without — the live accounts server, the only copy of anything: that does not belong alone on a single ageing desktop. Either it gets a second node, a proper server with redundant parts, or it stays on its own machine. The technology is free; the judgement is the part you cannot download.
See the full Proxmox installation and first VM creation, live and step by step.
▶️ Watch Full VideoWith backups on separate hardware and another Proxmox-capable PC available, restoring your key VMs is typically a same-day, often same-hour job. Without backups, the answer is 'rebuild everything from memory' — which is why the backup rule outranks every other precaution on this page.
It helps with planned maintenance and one hardware failure, but two nodes cannot reliably agree on which of them is still alive when they lose contact — the quorum problem. The usual fixes are a third tiny voting device or accepting the limits knowingly. For most small setups, tested backups on one host beat a fragile two-node cluster.
A fair and often skipped question. An ageing desktop can draw 60–120 watts idling — measure yours with a plug meter over a week and do the local tariff maths. Sometimes a newer low-power mini PC pays for itself; sometimes the old box is genuinely cheap to run. Measure, do not guess.