Operating System Images
Pre-configured operating system images used to create and deploy virtual machines.
An operating system (OS) image is a pre-configured representation of an operating system used for creating and deploying virtual machines (VMs). It includes all the necessary files, configurations, and settings needed to run the operating system in a virtualized environment. These OS images contain a fully installed and configured operating system, including system libraries, application software, and system settings essential for the functionality and performance of the virtual machine. By using OS images, administrators can quickly and consistently deploy multiple instances of virtual machines, ensuring uniformity and reducing the time required for installation and setup. Choose from our range of images, including those pre-installed with NVIDIA drivers and the CUDA toolkit, compatible with our GPU offerings.
Images Offered by Hyperstack
To select an Ubuntu image for a new virtual machine, simply choose your desired image from the dropdown menu in the "Choose an image" section of the "Deploy Virtual Machine” page. Explore details about each Ubuntu image provided by Hyperstack below.
Use the List Images API to get the latest and complete list of all images offered by Hyperstack.
Virtual machines and clusters that were previously created using older image versions (for example, Server 20.04) will continue to operate normally. These resources can still be hibernated and restored without issue. However, older images are no longer offered for new deployments, and we strongly recommend upgrading to one of the currently supported OS versions listed below to benefit from ongoing security updates, improved performance, and the latest platform features.
Driver/CUDA Pre-installed Images
These images come pre-configured with NVIDIA drivers and the CUDA toolkit, optimized for GPU workloads right out of the box. They are available with or without Docker support.
R535 Images
| Hyperstack Image Name | API Image Name | Distribution | NVIDIA Driver | CUDA Toolkit | Shipped With Kernel Version | Kernel Version Installed on Image | Regions |
|---|---|---|---|---|---|---|---|
| Server 22.04 LTS R535 CUDA 12.2 | Ubuntu Server 22.04 LTS R535 CUDA 12.2 | Ubuntu 22.04 | R535 | 12.2 | 5.15 (GA) | 6.5 (HWE) | CANADA-1, NORWAY-1, US-1 |
| Server 22.04 LTS R535 CUDA 12.2 with Docker | Ubuntu Server 22.04 LTS R535 CUDA 12.2 with Docker | Ubuntu 22.04 | R535 | 12.2 | 5.15 (GA) | 6.5 (HWE) | CANADA-1, NORWAY-1, US-1 |
R550 Images
| Hyperstack Image Name | API Image Name | Distribution | NVIDIA Driver | CUDA Toolkit | Shipped With Kernel Version | Kernel Version Installed on Image | Regions |
|---|---|---|---|---|---|---|---|
| Server 22.04 LTS R550 CUDA 12.4 | Ubuntu Server 22.04 LTS R550 CUDA 12.4 | Ubuntu 22.04 | R550 | 12.4 | 5.15 (GA) | 6.8 (HWE) | CANADA-1, NORWAY-1, US-1 |
| Server 22.04 LTS R550 CUDA 12.4 with Docker | Ubuntu Server 22.04 LTS R550 CUDA 12.4 with Docker | Ubuntu 22.04 | R550 | 12.4 | 5.15 (GA) | 6.8 (HWE) | CANADA-1, NORWAY-1, US-1 |
R570 Images
| Hyperstack Image Name | API Image Name | Distribution | NVIDIA Driver | CUDA Toolkit | Shipped With Kernel Version | Kernel Version Installed on Image | Regions |
|---|---|---|---|---|---|---|---|
| Server 22.04 LTS R570 CUDA 12.8 | Ubuntu Server 22.04 LTS R570 CUDA 12.8 | Ubuntu 22.04 | R570 | 12.8 | 5.15 (GA) | 6.9 (HWE) | CANADA-1, NORWAY-1, US-1 |
| Server 22.04 LTS R570 CUDA 12.8 with Docker | Ubuntu Server 22.04 LTS R570 CUDA 12.8 with Docker | Ubuntu 22.04 | R570 | 12.8 | 5.15 (GA) | 6.9 (HWE) | CANADA-1, NORWAY-1, US-1 |
| Server 24.04 LTS R570 CUDA 12.8 | Ubuntu Server 24.04 LTS R570 CUDA 12.8 | Ubuntu 24.04 | R570 | 12.8 | 6.8 (GA) | 6.8 (GA) | CANADA-1, NORWAY-1, US-1 |
| Server 24.04 LTS R570 CUDA 12.8 with Docker | Ubuntu Server 24.04 LTS R570 CUDA 12.8 with Docker | Ubuntu 24.04 | R570 | 12.8 | 6.8 (GA) | 6.8 (GA) | CANADA-1, NORWAY-1, US-1 |
R595 Images
This image is required for the B300 flavor, which supports no other image.
| Hyperstack Image Name | API Image Name | Distribution | NVIDIA Driver | CUDA Toolkit | Regions |
|---|---|---|---|---|---|
| Server 24.04 LTS R595 CUDA 13.2 with Docker | Ubuntu Server 24.04 LTS R595 CUDA 13.2 with Docker | Ubuntu 24.04 | R595 | 13.2 | CANADA-1 |
Vanilla Images (No Driver/CUDA Installed)
These images are minimal Ubuntu installations without NVIDIA drivers or CUDA, suitable for standard Linux workloads or custom installations. They are ideal for CPU-based tasks or user-managed GPU installations.
| Hyperstack Image Name | API Image Name | Distribution | NVIDIA Driver | CUDA Toolkit | Shipped With Kernel Version | Kernel Version Installed on Image | Regions |
|---|---|---|---|---|---|---|---|
| Server 22.04 LTS | Ubuntu Server 22.04 LTS (Jammy Jellyfish) | Ubuntu 22.04 | Not installed | Not installed | 5.15 (GA) | 6.8 (HWE) | CANADA-1, NORWAY-1, US-1 |
| Server 24.04 LTS | Ubuntu Server 24.04 LTS (Noble Numbat) | Ubuntu 24.04 | Not installed | Not installed | 6.8 (GA) | 6.8 (GA) | CANADA-1, NORWAY-1, US-1 |
AlmaLinux and Debian Images
These plain images do not include GPU drivers or CUDA and are ideal for lightweight or general-purpose workloads.
| Hyperstack Image Name | API Image Name | Distribution | Version | Size | Regions |
|---|---|---|---|---|---|
| AlmaLinux 9 Plain | AlmaLinux 9 Plain | AlmaLinux | 9 Plain | 1.46 GiB | CANADA-1, NORWAY-1, US-1 |
| Debian 12 Plain | Debian 12 Plain | Debian | 12 Plain | 1.92 GiB | CANADA-1, NORWAY-1, US-1 |
Kubernetes Compatible Images
The image listed below is compatible for deploying Kubernetes clusters in Hyperstack. Kubernetes-compatible images are marked with the k8s label, which is visible in the Hyperstack UI and returned by the List Images API.
| Hyperstack Image Name | API Image Name | Distribution | NVIDIA Driver | CUDA Toolkit | Shipped With Kernel Version | Kernel Version Installed on Image | Regions |
|---|---|---|---|---|---|---|---|
| Server 22.04 LTS R535 CUDA 12.2 | Ubuntu Server 22.04 LTS R535 CUDA 12.2 | Ubuntu 22.04 | R535 | 12.2 | 5.15 (GA) | 6.5 (HWE) | CANADA-1, NORWAY-1, US-1 |
Image and Flavor Compatibility for VM Deployment
Not every image runs on every flavor. To help guide successful VM deployments and improve performance outcomes, Hyperstack checks the selected image and flavor against any restrictions defined for that pair before launch, and displays an alert on the Virtual Machine Deployment page when the two are not compatible.
A restriction can be defined on either side of the pair, and each side is evaluated independently:
- An image restricted to flavors: the image runs only on the flavors it is approved for.
- A flavor restricted to images: the flavor launches only the images it is approved for. The image itself stays generally available, so other flavors can still use it.
Where a restriction applies, it either warns or blocks:
- Compatibility Warning: The pairing may reduce performance. Deployment is allowed, and an alert is shown.
- Incompatible Selection: The pairing is not supported. Deployment is prevented, and the alert explains what to select instead.
Each side reports only its own restrictions:
- Call the List Images API and read the
flavor_restrictionsobject on an image to see the flavors it is approved for. - Call the List Flavors API and read the
image_restrictionsobject on a flavor to see the images it is approved for.
A flavor's image restrictions never appear in flavor_restrictions, so check both objects before deploying.
If a virtual machine is created with an incompatible pair, the request fails with the error_reason compatibility_blocked and returns the approved set in suggested_images or suggested_flavors, so you can retry with a supported combination.