Getting started with RHEL for Workstations on AWS

Maybe you have a high-end workstation with a top-of-the-line GPU, but now the world is opening up to travel again. Are you planning on taking your desktop as a carry-on?

Maybe your graphical workstation is aging out and the cost of replacing all the hardware seems daunting. About the time you have replaced all the components, you realize that your motherboard can't support all your needs.

A way to avoid issues like these is to turn to the cloud. Red Hat and AWS have collaborated to bring Red Hat Enterprise Linux for Workstations to the Amazon Marketplace.

What is Red Hat Enterprise Linux for Workstations?

People love RHEL for its servers because they know exactly what they are getting in terms of quality and longevity. Why should your desktop environment be any different? 

RHEL for Workstations brings high-quality graphics hardware, like NVIDIA GPUs, to the workstation. Our workstation offering includes all of the same deployment options as our Server offering: bare metal, virtual and private or public cloud.

RHEL for Workstations also gets the same 10-year lifecycle and enterprise support. Imagine building a robust, scientific-compute rig and knowing that it will be supported for a decade! Not only that, but our product comes with an ever-growing list of certified hardware and partner-built software.

Starting with RHEL 9, we've upgraded our RHEL for Workstations offering to GNOME 40.

Couple some professional-grade hardware, a 10-year lifecycle, the Linux kernel and a great desktop environment, and you'll have an excellent workstation to get your work done. 

How does RHEL for Cloud Workstations work?

To start deploying RHEL for Workstations, you'll need to log into your AWS account, ensure you are connected to your desired data center and not hit any vCPU quota limits.

Next, you need to decide which driver makes the most sense for you:

• Tesla drivers
  • These drivers are intended primarily for compute workloads, which use GPUs for computational tasks such as parallelized floating-point calculations for machine learning and fast Fourier transforms for high performance computing applications.
• GRID drivers
  • These drivers are certified to provide optimal performance for professional visualization applications that render content such as 3D models or high-resolution videos. You can configure GRID drivers to support two modes. Quadro Virtual Workstations provide access to four 4K displays per GPU. GRID vApps provide RDSH App hosting capabilities.

Source: Install NVIDIA drivers on Linux instances

The easiest way to spin up a RHEL cloud workstation is by searching the Amazon Marketplace Image (AMI) catalog. Navigate to EC2, then select the catalog under images.

Once in the catalog, select "AWS Marketplace AMIs" and type RHEL GRID in the search box. This will bring up the latest version of the image.

Review the price, version and support agreement then hit continue to move forward. This will return you to the catalog. Click "Launch Instance with AMI".

We are ready to configure our new RHEL instance. Let's give our system a name and verify we have the correct AMI (just in case).

You can accept the defaults for type, network and storage, but make sure you pick a key pair to enable easy connection to the remote machine.

That's all the configuration we need to kick off our EC2 instance's build. Click "Launch instance" and give the system a few minutes to spin up and go through the AWS system checks.

How do I connect to my cloud workstation?

Bear in mind through this section that this is a new offering and is still under active development. There are several settings and configurations that you may want to address before connecting to your RHEL workstation for the first time:

First, we need to view our instance settings and modify the security group. There need to be three settings under inbound rules.

We need to click "Edit inbound rules". There should be SSH available on port 22 and port 8443 for both TCP and UDP. (At the time of this writing, 8443/UDP is not part of the default security group.)

Save these changes and return to the instance summary. Across the top bar, click on Connect and follow the instructions for the SSH client. (Remember the keypair we configured earlier?)

The following steps will get easier as newer marketplace images are released. For now, let's walk through the commands needed to get the best connectivity out of your cloud instance:

sudo dnf remove $(sudo dnf list installed | grep '@cuda' | awk '{ print $1 }') -y # Remove existing drivers
sudo rm -f /etc/yum.repos.d/cuda-rhel8.repo # Remove the cuda repo
sudo dnf upgrade -y # Update the system packages
sudo dnf install -y https://dl.fedoraproject.org/pub/epel/epel-release-latest-8.noarch.rpm # Configure the EPEL repository
sudo dnf install -y make gcc elfutils-libelf-devel libglvnd-devel kernel-devel-$(uname -r) dkms # Install build dependencies
sudo dnf install -y @workstation-product-environment # Install workstation packages
sudo reboot # Reboot to clear out the old drivers

curl "https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip" -o "awscliv2.zip" # Download the AWS CLI utility
unzip awscliv2.zip # open the archive
sudo ./aws/install # install AWS CLI
aws configure # setup your AWS account
aws s3 cp --recursive s3://ec2-linux-nvidia-drivers/latest/ . # Download the latest NVIDIA driver
chmod +x NVIDIA-Linux-x86_64*.run # add execute permissions to the installer
sudo /bin/sh ./NVIDIA-Linux-x86_64*.run # run the installer
sudo reboot

echo "options nvidia NVreg_EnableGpuFirmware=0" | sudo tee --append /etc/modprobe.d/nvidia.conf # configure the NVIDIA kernel parameters
sudo nvidia-xconfig --preserve-busid --enable-all-gpus --connected-monitor='DFP-0,DFP-1' --flatpanel-properties="Dithering = Disabled" # Disable dithering

sudo firewall-cmd --zone public --add-port 8443/udp # Add 8443/UDP to the firewall
sudo firewall-cmd --runtime-to-permanent # save the firewall configuration

sudo sed -i s/^\#create-session/create-session/ /etc/dcv/dcv.conf # set DCV to create a new session
sudo sed -i s/^'#owner = ""'/'owner = "ec2-user"'/ /etc/dcv/dcv.conf # configure DCV to user (ec2-user in this example)
sudo sed -i s/^\#authentication\=\"none\"/authentication\=\“system\”/ /etc/dcv/dcv.conf # set DCV authentication to system
sudo reboot # one final reboot

I know that was a lot, but it will be worth it! On your local system, you'll need to install Nice DCV. It is cross-platform and supports Mac, Windows and Linux clients.

With the DCV View installed, input the public address of your remote workstation.

Put your system user and password into the login prompt, and there you go! When your connection completes, you'll be looking at the login screen of your brand-new RHEL cloud workstation!

Wrap up

Whether you are a graphics designer, animator, scientific researcher or architect, Red Hat Enterprise Linux for Workstations is the way to go to bring all the stability and security you are used to with RHEL for Server to your cloud-based desktop. That way, no matter where you are or what hardware you have with you, you will always have access to a GPU-enabled Workstation.

Get started

• Get a trial for Red Hat Enterprise Linux
• Sign up for a no-cost RHEL Developer Subscription