A Proxmox VE home lab is one spare machine that runs several virtual machines and Linux containers, managed from a web browser. This guide covers the decisions a program download page does not: how much hardware you actually need, whether to pick the default ext4 with LVM-thin or ZFS before you install, and how to create your first VM, your first LXC container and get both on your network. For install internals, the repository switch and clustering, see the Proxmox VE page. Facts were checked on 18 November 2026 against the Proxmox roadmap, the requirements page and pve-docs. Not tested yet: we did not install these builds on lab hardware for this article, so treat versions and commands as a starting point and read each screen before you confirm it.
What you need and what was checked on 18 November 2026
| Item | Detail |
|---|---|
| Proxmox VE | Free, open source (GNU AGPL-3.0). The roadmap lists 9.2 as the current release; the x86-64 ISO is 9.2-1 (21 May 2026), the Arm64 ISO 9.2-1-arm64 (5 August 2026) |
| CPU | 64-bit Intel/AMD with the VT-x or AMD-V flag, or 64-bit Arm (Armv9-A or newer). Enable virtualization in the BIOS first |
| Memory (official) | Minimum 1 GB for testing only; 2 GB for the host and Proxmox services, plus memory for each guest. ZFS or Ceph add roughly 1 GB of RAM per TB of used storage |
| Disk and network | The requirements page recommends fast, redundant storage (SSD) and redundant Gbit NICs; a single SSD and one NIC are fine for a lab |
| Access | Web interface on port 8006 from any modern browser; no client software to install |
The 1 GB minimum boots the hypervisor but leaves nothing for guests; size the host for the guests you plan to run.
Pick your storage before you install
The installer asks for a filesystem on the target disk and the answer is hard to change later, so decide first. The choices are ext4 (the default), XFS, ZFS and BTRFS, which is still a technology preview. The practical split for a single home-lab box is ext4 with LVM-thin versus ZFS.
| Criterion | ext4 + LVM-thin (default) | ZFS |
|---|---|---|
| Host RAM | Light | Uses an ARC cache in RAM, which new installs cap at 10% of host memory, up to 16 GiB |
| Disks | One disk is fine | At its best with two or more for a mirror or RAIDZ |
| Snapshots | Yes (thin volumes) | Yes (native) |
| Integrity and compression | No | Yes — continuous integrity checking and built-in compression |
| Default storages created | local and local-lvm | local and local-zfs |
On one disk with limited RAM, the default ext4 with LVM-thin is the lower-overhead choice and still gives you snapshots. Pick ZFS when you have two or more disks or you specifically want checksums and compression and have the RAM to spare. ECC memory is not listed as a requirement either way; it protects data integrity but Proxmox does not demand it.
Whichever you choose, the install leaves you two storages. local is a Directory storage at /var/lib/vz that holds ISO images, container templates and backups. local-lvm (or local-zfs) holds the VM disk images and container root filesystems.
Install Proxmox VE
Write the ISO to a USB stick and boot from it. The installer asks for the target disk and filesystem (use the Options button to change ext4 to ZFS), a root password, an e-mail address and the management IP, hostname and gateway. When it finishes, open https://<server-ip>:8006 in a browser and log in as root with the Linux PAM realm. Expect a self-signed certificate warning.
A fresh install points at the enterprise update repository, which needs a paid key. For a home lab, switch to the free pve-no-subscription repository under Node > Updates > Repositories before you run updates; the Proxmox VE page shows the exact repository file. The licence-key notice at login is informational and does not lock any feature.
Create your first VM
A virtual machine (KVM/QEMU) emulates a full computer, so it can run Windows, BSD or any Linux kernel. Start with a Linux server VM:
- Upload an ISO: select the node, open
local> ISO Images and use Upload, or Download from URL to fetch it on the server. - Click Create VM. On the OS tab pick the ISO. Leave the defaults on System for a Linux guest.
- Put the disk on
local-lvmand keep the VirtIO SCSI single controller and a VirtIO network card, so the guest uses the fast paravirtual drivers. - On the CPU tab, set Type to
hostso the guest gets the host’s CPU flags; the Proxmox docs note this gives the best performance when you are not live-migrating between different CPUs. - Give it memory, attach it to bridge
vmbr0, finish, then Start and open the Console to run the installer.
After the OS is installed, install the qemu-guest-agent package inside the guest and tick QEMU Guest Agent in the VM’s Options. The web UI then reports the guest IP, clean shutdowns work, and backups are consistent. QEMU is the emulator doing the work under every Proxmox VM.
Create your first LXC container
A container is lighter than a VM: it shares the host kernel instead of emulating a whole machine, so it starts in seconds and uses little memory. That makes LXC a good fit for always-on Linux services in a home lab. You cannot run Windows or a different kernel in one, though.
Containers start from a template you download first. From the node shell or the Shell button:
pveam update
pveam available --section system
pveam download local <template-name>Use a template name shown by pveam available — a current Debian or Ubuntu standard image, for example. Then create the container in the UI with Create CT: pick the template from local, put the root disk on local-lvm, give it memory and a password, and attach it to vmbr0. The pct command-line tool does the same job from the shell if you prefer.
Get your guests on the network
The installer creates a single Linux bridge named vmbr0, connected to the management interface you chose during setup. VMs and containers attached to vmbr0 behave as if they were plugged straight into your physical network, so they appear on your home LAN and pick up an address from your router’s DHCP. That is all most labs need. For an isolated network, add a second bridge with no physical port and attach the test guests to it, or use VLAN tags on vmbr0.
What a single node can’t do
- One node has no live migration and no high availability: when the host reboots, every guest goes down with it. High availability and cluster quorum need at least three nodes (or two plus a QDevice).
- A single disk is a single point of failure. ZFS on one disk gives you checksums and snapshots but no redundancy — if the disk dies, the pool dies. Back up to a separate disk or a NAS regardless of filesystem.
- The free repository receives packages before the enterprise one; test an update before you trust it on anything that matters.
Who a Proxmox home lab suits
| Situation | Fit |
|---|---|
| Learning virtualization on one mini PC or a spare desktop | Good: a single box runs several VMs and containers side by side |
| A few always-on home services (DNS, dashboards, a media stack) | Good: LXC containers keep the memory and CPU cost low |
| You need redundancy or zero downtime | Not from one node; plan three nodes or accept that a reboot stops everything |
| You just want a throwaway VM on your laptop | Overkill; a desktop hypervisor such as VMware Workstation Pro or Vagrant is simpler |
When you find yourself rebuilding the same VMs by hand, describe them in code instead: see OpenTofu + Proxmox VE. The virtualization and container tools hub lists the wider set of options.
FAQ
How much RAM do I need for a Proxmox VE home lab?
The official minimum is 1 GB, but that is for testing only. Budget 2 GB for the host plus the memory each guest needs, and roughly 1 GB per TB of used storage if you run ZFS. As a rough rule of thumb, around 8 GB is comfortable for several light Linux services, with more for Windows VMs.
Should I choose ZFS or LVM-thin on a single disk?
On one disk with limited RAM, the default ext4 with LVM-thin is lighter and still supports snapshots. ZFS adds integrity checking and compression, but it caches in RAM and gives no redundancy on a single disk, so the extra cost mainly buys data-integrity checking.
Can I run Proxmox VE on a mini PC or an old desktop?
Yes, as long as the CPU is 64-bit with the VT-x or AMD-V flag. Enable virtualization in the BIOS or UEFI first. A mini PC with an SSD and enough RAM for your guests makes a quiet, capable lab.
Does Proxmox VE need ECC RAM?
No. The official requirements page lists no ECC requirement. ECC memory reduces the risk of silent data corruption and is worth having for storage you care about, but Proxmox and ZFS both run without it.
Should I use a VM or an LXC container?
Use a container for a Linux service where sharing the host kernel is fine; it is lighter and faster to start. Use a VM when you need Windows, a different kernel, kernel modules or stronger isolation between the guest and the host.
How do my VMs get on the network?
By default they attach to the vmbr0 bridge, which is connected to the host’s management network interface. Guests on it appear on your LAN and get an IP from your router, exactly as a physical machine would. Add another bridge or VLAN tags when you want separate networks.
Last updated: 18 November 2026 · ITForgePro editorial team. Licence, version and platform details are checked against each developer's official documentation.





