Proxmox VE – Initial Installation & Setup Guide

Proxmox VE is the platform that turned a repurposed compact PC into the main virtualisation host for Realm Labs.

The host is a Shuttle DS81 named QuanChi. It is not enterprise server hardware and it was never intended to be. The goal was to build a small, quiet and understandable hypervisor that could run Linux containers, virtual machines, monitoring services and internal tools without leaving a full-size server running around the clock.

This guide covers the complete installation process, but it also explains the decisions behind it: why I reused the Shuttle, why the management address is static, why storage is separated from compute, and why I treat third-party post-install scripts as optional rather than essential.

What This Build Is Designed to Do

QuanChi provides the compute layer for Realm Labs. It runs the workloads, but it is not expected to be the only home of their important data.

• Run lightweight Linux containers and selected virtual machines
• Keep services separated instead of installing everything on one operating system
• Provide a simple web interface for resource, storage and network management
• Connect to Seido, the Synology DS224+, for network storage and backups
• Remain small enough that rebuilding the host is realistic rather than terrifying

The wider design is covered in Proxmox VE: The Infrastructure Foundation for Realm Labs. This article concentrates on getting the physical host from bare metal to a stable starting point.

Hardware Used in This Build

The exact Shuttle DS81 and HP switches are older hardware, so the purchasing links below point to the current equivalents already listed on the Realm Labs hardware page.

Affiliate disclosure: The Amazon buttons are affiliate links. Realm Labs may receive a small commission from a qualifying purchase at no additional cost to you.

Why the Shuttle DS81 Is Enough

It is easy to assume that a useful Proxmox host needs server-grade processors, huge amounts of memory and a rack full of disks. That depends entirely on the workloads.

QuanChi is intended mainly for infrastructure services and Linux containers. An LXC does not carry the same overhead as a complete virtual machine, so 16 GB of memory can go surprisingly far when services are sized sensibly. The compact Shuttle is also quiet, economical to leave running and easy to place in the garage lab.

The limitation is equally important: this is a single host, not a high-availability cluster. Maintenance or hardware failure means downtime. I accept that trade-off because the services are documented, backups live elsewhere and the host can be rebuilt.

The Installation Plan

The installation itself is straightforward. Most avoidable problems come from decisions made before clicking Install: choosing an inappropriate disk layout, using a temporary address, entering a hostname that DNS cannot resolve, or treating local storage as the only copy of important data.

Step 1: Download and Verify Proxmox VE

Download the current ISO from the official Proxmox VE download page. At the time of this September 2026 update, Proxmox VE 9.2 is the current x86-64 installer.

The download page publishes a SHA-256 checksum. Verify it before writing the USB drive; this confirms that the image arrived intact and is the file Proxmox published.

Get-FileHash .\proxmox-ve_*.iso -Algorithm SHA256

On Linux, the equivalent check is:

sha256sum proxmox-ve_*.iso

Compare the complete result with the value on the official download page. Do not continue if it differs.

Step 2: Create the Installer USB

Write the ISO to an 8 GB or larger USB drive using a tool such as Rufus, Balena Etcher or Ventoy. If Rufus asks whether to use ISO or DD mode, DD mode is the most direct representation of the installer image.

This process erases the selected USB drive. Check the device name and capacity twice before starting; installation media is not a good place for guesswork.

Step 3: Prepare the Shuttle Firmware

With the installer inserted, start the Shuttle and enter its firmware setup. On the DS81 this is normally reached with F2 or Delete during startup.

• Enable Intel VT-x so the processor can provide hardware-assisted virtualisation.
• Enable VT-d if PCI or device passthrough may be used later.
• Confirm that the internal system disk and wired network interface are detected.
• Select the USB device as the temporary boot target.

I use wired Ethernet for a hypervisor. Wi-Fi is convenient for clients, but storage traffic, backups and management access all benefit from the predictable connection provided by the Realm Labs wired network.

Step 4: Work Through the Installer

Choose the Target Disk

Select the Shuttle’s internal system drive. This is destructive: the selected disk will be repartitioned and its existing data removed.

For a compact single-drive node, a straightforward local filesystem is easier to justify than designing around ZFS without redundant disks. ZFS is excellent when its features match the hardware and recovery plan, but choosing it simply because it appears more advanced does not create redundancy.

Set Location and Time

Select the correct country, keyboard and time zone. Accurate time matters for logs, scheduled backups, authentication and correlating events between Proxmox, Synology and monitoring systems.

Create the Administrator Login

Use a unique root password and a working email address for system notifications. I do not publish real credentials, recovery codes or private management addresses in screenshots or examples.

Configure Networking

The management interface needs a stable identity. In Realm Labs the host is known as QuanChi, and the domain infrastructure is provided through Seido and order.realm.

Hostname:  quanchi.order.realm
Address:   192.168.68.50/24  (illustrative)
Gateway:   your router address
DNS:       the DNS server used by your network

The example address shows the Realm Labs subnet but is not presented as QuanChi’s published management address. Choose an unused address, reserve or document it, and make sure the hostname resolves consistently.

This matters later. NFS storage definitions, monitoring checks, browser bookmarks and other systems all need a predictable destination. The lessons from changing infrastructure addresses are covered in Changing a Proxmox NFS Server IP.

Step 5: First Login

After installation, remove the USB drive and allow the Shuttle to boot from its internal disk. The console displays the management URL:

https://your-proxmox-address:8006/

• Open the address from a machine on the trusted local network.
• Review the warning for the initial self-signed certificate.
• Sign in as root using the Linux PAM authentication realm.
• Confirm the node name, storage and network bridge before creating workloads.

The certificate warning is expected on a new private installation, but it should not become an excuse to ignore every browser security warning. Confirm that the address and certificate belong to the host you just installed.

Step 6: Repositories and Updates

Proxmox offers an enterprise repository for supported subscriptions and a no-subscription repository for community use. Pick the repository that matches how the host will be supported, then update the package lists and install available updates before deploying workloads.

apt update
apt full-upgrade

Repository configuration can be managed from the Proxmox interface. This is the clearest route because you can see exactly which sources are enabled.

Optional Community Post-Install Script

The maintained Proxmox VE Helper-Scripts project also provides a PVE Post Install script. It can manage repositories, the subscription notice, updates and related housekeeping.

bash -c "$(curl -fsSL https://raw.githubusercontent.com/community-scripts/ProxmoxVE/main/tools/pve/post-pve-install.sh)"

This is a community-maintained script, not a requirement for Proxmox to work. Read the current script and its documented options before piping it into a root shell. Paths and behaviour can change, which is exactly why an old command should not be copied blindly from an old guide.

Step 7: Connect Storage Without Confusing It with Backup

Once QuanChi is stable, network storage can be added from Seido. In Realm Labs, Proxmox provides compute while the Synology provides shared storage and separate backup capacity.

This separation makes the host easier to recover. A container or VM can be recreated, and the compute host can be replaced, without automatically losing every piece of application data at the same time.

It does not mean that NFS is magically a backup. A mounted share is still exposed to mistakes and configuration problems. Keep independent, versioned copies of the data that matters. The full storage design is documented in Building the Storage Backbone of Realm Labs, while What I Back Up Before a Major Proxmox Upgrade covers the recovery-critical items.

Step 8: Prove the Basics Before Adding Workloads

• Reboot the host and confirm it returns without the installer.
• Resolve QuanChi by name and reach the management interface from the trusted network.
• Check that the system clock, gateway and DNS are correct.
• Confirm repository status and install outstanding updates.
• Create a small test container before migrating an important service.
• Test storage access and a restore path rather than assuming the backup is usable.

I prefer proving one layer at a time. If the host, network and storage are validated separately, later faults are much easier to isolate than when ten services are deployed at once.

What I Would Do Differently on a Larger Host

The Shuttle works because the design accepts its limits. A larger build would justify more memory, mirrored local drives, additional network interfaces and possibly ZFS. A second Proxmox node would also make maintenance less disruptive.

Those upgrades would solve different problems; they would not make the initial installation fundamentally better. Clear naming, stable networking, separated storage and tested recovery remain more valuable than buying impressive hardware without a plan.

The Result

QuanChi started as a repurposed compact PC. Installing Proxmox VE gave it a clear role as the compute foundation for Realm Labs.

The important part was not simply getting the installer to finish. It was deciding what the host should own, what should live on Seido, how the node would be found on the network, and how it could be recovered when something eventually changes.

That is the difference between a machine that happens to run virtual machines and infrastructure that can be maintained with confidence.