Skip to main content

Creating the Cluster

A Gallium Cluster requires three servers. This page covers booting your servers, installing the first one, and establishing the cluster.

You can boot and adopt all three servers upfront, or start with just the first server and add the remaining two later. Either way, only the first server is fully configured and installed during cluster creation — the others are added afterward from within the cluster (see Adding Hypervisors).

Step 1: Setup​

Download Boot Media​

  1. Log in to the Gallium Console.
  2. Navigate to Deployments → New Deployment → Cluster.
  3. Under Download Boot Media, choose one of the two available installers:

Full Installation ISO — A larger (~1 GB) image that contains everything needed to install without a network connection during boot. Use this if your server has no internet access at install time, or if network conditions make the iPXE download unreliable.

Network Boot ISO (Recommended) — A small (~4 MB) iPXE-based image that downloads the latest installer over the network at boot time. This is the recommended option for most users, particularly when mounting media via a remote management interface (iLO, iDRAC, or similar IPMI).

tip

We recommend the Network Boot ISO for most deployments. It's faster to download and transfer, works well over remote management consoles, and always installs the latest version.

Each download card in the Console shows the file size and a SHA256 checksum. Verify the checksum after downloading to confirm the image is intact.

Boot Your Servers​

  1. Create bootable media from the downloaded ISO — write it to a USB drive or DVD, or mount it directly via your server's remote management interface (iLO, iDRAC, or similar IPMI virtual media).
  2. Configure your server to boot from the installation media.
  3. Power on or reboot the server.
  4. The installer will start automatically. When it's ready, it will display an install code on screen.
tip

You can boot all three servers at this stage if they are ready. Each server will display its own install code. This lets you adopt them all upfront and skip booting new servers later when adding hypervisors.

info

Note down each install code — you'll enter it in the Gallium Console in the next step to link the server to your account.

Adopt Your Installations​

On the same page where you downloaded the installer, scroll down to the Adopt Your Installations section.

  • Enter the Install Code displayed on your server's screen.
  • Optionally, provide a Hypervisor Name to identify this server. If left blank, a default name will be assigned.

Click Adopt Server to proceed. The Console will connect to the server and retrieve its hardware details. Repeat this for each server you have booted.

info

You can adopt all three servers at this stage, but only one will be configured as the first hypervisor during cluster creation. Any additional adopted servers will appear as pending hypervisors that you can add to the cluster later — see Adding Hypervisors.

Select First Hypervisor​

Select which server to configure first from the list of adopted servers. The table shows each server's name, IP address, resources (CPU cores and memory), and status.

Click Continue to Configuration.

Step 2: Configure​

Deployment Name​

Enter a Deployment Name for the cluster. This is the name that will appear in the Gallium Console to identify this cluster deployment.

Installation Disk​

Select the disk where Gallium will be installed. The Console displays all eligible disks with their model, serial number, size, and bus type (SATA, NVMe, SAS).

Network Configuration​

Management Network Interfaces — Select 1–2 network interfaces for the management network. The Console shows each NIC's name, MAC address, model, carrier status (Up/Down), and link speed.

When two NICs are selected, they operate in an active/failover configuration. Use the swap button to change which NIC is Primary and which is Secondary.

NICs with carrier status "Down" cannot be selected.

Network Address — Configure the following:

  • Virtual IP — A shared IP address for the cluster. This address floats between cluster members and provides a stable endpoint for accessing the deployment, regardless of which server is currently active. Must be a valid IPv4 address on the same subnet as the management IPs (e.g., 192.168.1.50).
  • Management IP — The IP address for this specific server, in CIDR notation (e.g., 192.168.1.100/24).
  • Gateway — The default gateway address (e.g., 192.168.1.1).
info

Cluster network configuration uses static IPs only — there is no DHCP option. Each cluster member needs a unique management IP, and the cluster needs a shared Virtual IP.

Step 3: Confirm​

Resource Requirements​

The Console checks that your server meets the minimum hardware requirements:

  • Memory — 24 GB RAM
  • CPU — 12 threads

If your server does not meet these requirements, the Console will display a warning and prevent you from proceeding.

Disk Erasure Confirmation​

You must explicitly confirm erasure for each disk that will be affected by the installation. Check the Erase checkbox next to each listed disk. If the server has an existing installation, all conflicting disks will be flagged and must also be confirmed.

danger

Disk erasure cannot be undone. All data on confirmed disks will be permanently destroyed.

Click Start Installation to begin. The installer will run on the server and the Console will redirect you to a status page where you can monitor progress.

Next Steps​

Once the first server is installed and the cluster is online, add two more hypervisors to complete the three-node cluster. If you already booted and adopted additional servers during setup, they will be available as pending hypervisors ready to configure.