SSH into Arch Linux

PT1 - Setting up SSH

Needed to set up SSH on Arch Linux to access it from another computer for comfort purposes. its an ancient beast that has a gtx 1650, learning cuda and thought i’d give prime intellect a whirl.

On a fresh Arch install, the SSH server might not be running yet. We’ll need to grab it and tell it to start up:

1. Grab OpenSSH from the repositories

sudo pacman -S openssh

2. Fire up the service and make sure it starts every time you boot

sudo systemctl enable --now sshd

2. Finding Your Target

Now you need to know where to point your other computer. Since everything is on the same home network, just run:

ip addr show

Look through the list for your Wi-Fi or Ethernet connection. You’re looking for a number that looks like 10.0.0.70 (or something similar starting with 192.168). That’s your address!

4. The Handshake

Grab your other laptop (your “Client”), open up a terminal or PowerShell, and type this in:

ssh username@10.0.0.70

PT2 - Fixing the gui boot issue

I had an issue where on boot nvidia and intel drivers would conflict, so i wasnt able to boot into the gui. here’s how i fixed it:

1. The Diagnosis

Using my SSH session, I checked the logs for SDDM (the KDE login manager):

journalctl -u sddm | tail -n 50

The logs showed a Segmentation Fault in libgallium. This confirmed a driver conflict: the Intel integrated graphics and the NVIDIA dedicated card were failing to coordinate during the boot process.

2. The Solution: Early KMS

The fix required telling the Linux kernel to load the NVIDIA drivers as early as possible (Early Kernel Mode Setting).

Step A: Update the Bootloader

I’m using systemd-boot. I had to edit my boot entry in /boot/loader/entries/ and add a specific flag to the options line: options … nvidia-drm.modeset=1

This forces the NVIDIA driver to initialize correctly before the GUI attempts to start.

Step B: Configure mkinitcpio

I updated the kernel’s initial RAM disk configuration to include the NVIDIA modules:

sudo nano /etc/mkinitcpio.conf

I added the modules to the list: MODULES=(nvidia nvidia_modeset nvidia_uvm nvidia_drm)

Then, I regenerated the images:

sudo mkinitcpio -P

3. Syncing the Drivers

To be safe, I ensured the proprietary NVIDIA drivers were fully updated and synced with my current kernel:

sudo pacman -Syu nvidia nvidia-utils

4. The Resurrection

With the config files updated and the drivers synced, I issued the final command via SSH:

sudo reboot

The laptop whirred, the kernel loaded the NVIDIA modules early, and the KDE Plasma login screen finally appeared. By using SSH as a backdoor, I saved my installation without ever needing to re-install.