Skip to main content

Prerequisites

Before you start, ensure you have: Download the latest talosctl.

Download the image

Go to https://factory.talos.dev select Single Board Computers, select the version and select Turing RK1 from the options. Choose your desired extensions and fill in the kernel command line arguments if needed. Download the disk image and decompress it:

Boot options

You can boot Talos from:
  1. booting from eMMC
  2. booting from a USB or NVMe (requires a spi image on the eMMC)

Booting from eMMC

Flash the image to the eMMC and power on the node: (or use the WebUI of the Turing Pi 2)
Proceed to bootstrapping the node.

Booting from USB or NVMe

Requirements

To boot from USB or NVMe, flash a u-boot SPI image (part of the SBC overlay) to the eMMC.

Steps

Skip step 1 if you already installed your NVMe drive.
  1. If you have a USB to NVMe adapter, write Talos image to the USB drive:
  2. Install the NVMe drive in the Turing Pi 2 board. If the NVMe drive is/was already installed:
    • Flash the Turing RK1 variant of Ubuntu to the eMMC.
    • Boot into the Ubuntu image and write the Talos image directly to the NVMe drive:
  3. Find the latest sbc-rockchip overlay, download and extract the SBC overlay image:
    • Find the latest release tag of the sbc-rockchip repo.
    • Download the sbc overlay image and extract the SPI image:
  4. Flash the eMMC with the Talos raw image (even if Talos was previously installed): (or use the WebUI of the Turing Pi 2)
  5. Flash the SPI image to set the boot order and remove unnecessary partitions: (or use the WebUI of the Turing Pi 2)
Talos will now boot from the NVMe/USB and enter maintenance mode.

Bootstrapping the node

To monitor boot messages, run: (repeat)
Wait until instructions for bootstrapping appear. Follow the UART instructions to connect to the interactive installer:
Alternatively, generate and apply a configuration:
Copy your talosconfig to ~/.talos/config and fill in the node field with the IP address of the node and endpoints. Once applied, the cluster will form, and you can use kubectl.

Retrieve the kubeconfig

Retrieve the admin kubeconfig by running: