Prepare Appliance Nodes on Proxmox
Create the appliance targets using the boot artifact and network values displayed by Management. These steps begin at Download and boot the Talos artifact in Install the appliance. For Management, choose the ISO VM or Docker, Compose or Podman.
The public Talos download links include the boot images and checksums used by Management.
Before you begin
Use a Proxmox account with permission to upload ISO images and create and configure the required VMs. Have the selected noCloud ISO, its verified checksum and the node plan available. Allocate the resources for each appliance node.
Create each VM
-
Upload the verified noCloud ISO under Storage > ISO Images > Upload.
-
Select Create VM for each planned node. Set its name and select the uploaded ISO. Do not start it yet.
-
On System, select q35 for Machine and OVMF (UEFI) for BIOS. Add an EFI Disk on persistent storage and leave Pre-Enroll keys disabled.
-
Set CPU cores and memory to the planned values. Use at least 4 cores and 16384 MiB memory.
-
Add the 135 GB OS disk. Under Hardware > Add > Hard Disk, add the separate Database and Data disks. The recommended capacities are 200 GB and 500 GB respectively.
-
Connect
net0to the bridge for the shared node subnet and keep the adapter connected. -
If you selected the boot artifact with the QEMU guest agent, enable Options > QEMU Guest Agent > Use QEMU Guest Agent.
Keep the EFI disk separate from the three installation disk roles. Do not clone an already installed appliance to create new targets.
Configure Networking and Boot
-
Under Hardware, confirm that the downloaded noCloud ISO is attached as a CD/DVD drive. Attach it if needed.
-
Select Add > CloudInit Drive and choose its storage.
-
Open Cloud-Init. Set DNS servers and the IP Config for
net0using that node’s values on the Management page: static IPv4 address and prefix, gateway and DNS servers. -
Select Regenerate Image.
-
Under Options > Boot Order, put the intended 135 GB OS disk first and the noCloud CD/DVD drive after it. Start the VM.
-
Repeat for every node, using each node’s own address and identity.
On the first boot, the OS disk is empty, so the VM continues to the ISO. After installation, the installed OS disk remains first. This boot order differs from the Management ISO VM, which always boots its Management ISO.
Verify and Continue
Confirm that all targets boot into maintenance mode and show their intended network addresses. Return to Management and select Start target discovery. Compare the observed VM identities before confirming targets or disks.
If discovery does not find a target, inspect the VM console, Cloud-Init values, generated image, network bridge and boot order. If Management rejects resources or boot mode, correct the VM configuration before retrying discovery. Continue with Confirm targets and install.