🛠 Network and storage setup on Xerxes Nano
Objectives
This tutorial guides beginners through the basic configuration of the Xerxes Nano board. The goal is to connect the board to a local network via its Gigabit Ethernet port and initialize M.2 storage. These steps are essential to transform the compute module (Jetson Orin Nano or Turing RK1) into a lightweight server or autonomous data processing node. We will focus on physical connections and minimal software checks.
Required Materials
- Xerxes Nano carrier board
- Compatible compute module (NVIDIA Jetson Orin Nano or Turing RK1)
- USB-C power supply or PoE injector
- RJ45 Ethernet cable
- NVMe M.2 SSD (M.2 format, PCIe Gen 3 compatible)
- Host computer with SSH terminal or connected screen/keyboard
Steps
- Module Installation : Gently insert the compute module into the Xerxes Nano socket. Ensure connectors are aligned before pressing firmly.
- Storage Mounting : Slide the M.2 SSD into the slot located under the module. Secure it with the provided screw. Note that this slot supports up to three PCIe lanes, offering high bandwidth.
- Network Connection : Plug the Ethernet cable into the RJ45 port. If using PoE power, ensure the power selection jumper is set to "PoE". Otherwise, use the USB-C port with the jumper set to "USB".
- Boot : Connect power. The fan (if connected to the PWM connector) should spin. Wait about 30 seconds for system startup.
- System Access : Log in via SSH or screen. Open a terminal to check devices.
- Storage Verification : Type the following command to list disks :
lsblk
Identify your SSD (often named nvme0n1).
- Formatting and Mounting : Format the drive to ext4 and create a mount point :
sudo mkfs.ext4 /dev/nvme0n1
sudo mkdir /mnt/data
sudo mount /dev/nvme0n1 /mnt/data
Verification
Run df -h to confirm storage is mounted and available. For network, type ip a to verify eth0 has a valid IP address. A ping to an external address (e.g., ping 8.8.8.8) confirms internet connectivity.
Going Further
Explore using the SFP+ cage for high-speed fiber connections. You can also test the OCuLink connector to add an external GPU if your module supports it, boosting AI computing power.