Deploying and Installing SUSE AI | SUSE AI 1.0
|Index|Deploying and Installing SUSE AI
On this pageSUSE AI 1.0<br>Deploying and Installing SUSE AI
WHAT?SUSE AI is a multi-layered enterprise AI solution that consists of several software components and dependencies.<br>WHY?Use this guide to deploy and install SUSE AI and its dependencies in a consistent and supported manner.<br>EFFORTIt takes less than one hour of reading time and requires advanced knowledge of Linux deployment.<br>GOALBy the end of this guide, you will have SUSE AI and all required dependencies installed and ready for use.
Revision History: Deploying and Installing SUSE AI
Publication Date: 2026-08-13
1 Installation overview<br>2 Installing the Linux and Kubernetes distribution2.1 Installing SUSE Linux Enterprise Server<br>2.2 Installing NVIDIA GPU drivers<br>2.3 Installing SUSE Rancher Prime: RKE2
3 Preparing the cluster for AI Library3.1 Installing SUSE Rancher Prime on a Kubernetes cluster<br>3.2 Installing the NVIDIA GPU Operator on the SUSE Rancher Prime: RKE2 cluster<br>3.3 Registering existing clusters<br>3.4 Assigning GPU nodes to applications<br>3.5 Installing SUSE Security<br>3.6 Setting up SUSE Observability for SUSE AI
4 Installing applications from AI Library4.1 What is SUSE Application Collection?<br>4.2 What is SUSE Registry?<br>4.3 Installation procedure<br>4.4 Installing cert-manager<br>4.5 Installing OpenSearch<br>4.6 Installing Milvus<br>4.7 Installing Ollama<br>4.8 Installing Open WebUI<br>4.9 Installing vLLM<br>4.10 Installing mcpo<br>4.11 Installing PyTorch<br>4.12 Installing Qdrant<br>4.13 Installing LiteLLM<br>4.14 Installing MLflow<br>4.15 Installing Kubeflow<br>4.16 Verifying SUSE AI Library applications
5 Alternative deployment5.1 Node installer
A Glossary<br>B Copyright<br>C GNU Free Documentation LicenseC1 0. PREAMBLE<br>C2 1. APPLICABILITY AND DEFINITIONS<br>C3 2. VERBATIM COPYING<br>C4 3. COPYING IN QUANTITY<br>C5 4. MODIFICATIONS<br>C6 5. COMBINING DOCUMENTS<br>C7 6. COLLECTIONS OF DOCUMENTS<br>C8 7. AGGREGATION WITH INDEPENDENT WORKS<br>C9 8. TRANSLATION<br>C10 9. TERMINATION<br>C11 1. FUTURE REVISIONS OF THIS LICENSE<br>C12 ADDENDUM: How to use this License for your documents
List of Figures<br>1.1 SUSE AI installation process<br>2.1 Language, keyboard and product selection<br>2.2 License agreement<br>2.3 Network settings<br>2.4 Registration<br>2.5 Extension and module selection<br>2.6 Add-on product<br>2.7 System role<br>2.8 Suggested partitioning<br>2.9 Clock and time zone<br>2.10 Local user creation<br>2.11 Password for the system administrator root<br>2.12 Installation settings<br>2.13 Confirm installation<br>2.14 Performing the installation<br>3.1 Rancher extensions<br>3.2 Enable extensions<br>3.3 Install UI extension<br>3.4 Reload extension<br>3.5 Install SUSE Security application<br>3.6 SUSE Security dasboard<br>3.7 Uninstalling extension<br>3.8 Rancher chart<br>3.9 Neuvector values<br>3.10 Neuvector deployed<br>3.11 Neuvector console access<br>3.12 Neuvector administrator users<br>3.13 Local admin<br>3.14 Rancher SSO<br>3.15 High-level overview of the SUSE Observability setup<br>3.16 Separate clusters for SUSE AI and SUSE Observability<br>3.17 New GenAI Observability menu item<br>4.1 Milvus page in the SUSE Application Collection<br>4.2 Adding a vLLM connection to Open WebUI<br>4.3 MLflow Web UI
List of Tables<br>4.1 Override file options for the Ollama Helm chart<br>4.2 Available options for the Open WebUI Helm chart<br>4.3 Available options for the PyTorch Helm chart<br>4.4 Options for the LiteLLM Helm chart
List of Examples<br>3.1 Click to Expand<br>4.1 Ollama Helm chart version 0.x.x<br>4.2 Ollama Helm chart version 1.x.x<br>4.3 Basic override file with GPU and two models pulled at startup<br>4.4 Basic override file with Ingress and no GPU<br>4.5 Open WebUI override file with Ollama included<br>4.6 Open WebUI override file with Ollama installed separately<br>4.7 Open WebUI override file with pipelines enabled<br>4.8 Open WebUI override file with a connection to vLLM<br>4.9 Stand-alone deployment of open-webui-pipelines<br>4.10 Minimal configuration<br>4.11 Validating the installation<br>4.12 Basic configuration<br>4.13 Loading prefetched models from persistent storage<br>4.14 Configuration with multiple models<br>4.15 CPU offloading<br>4.16 Shared remote KV cache storage with LMCache<br>4.17 Basic override file with GPU enabled<br>4.18 ConfigMap-based upload<br>4.19 Host-folder with files baked into the chart<br>4.20 Git repository clone: public with no authentication<br>4.21 Git repository clone: private with authentication<br>4.22 Basic override file when the cluster has no default storage class set.<br>4.23 An example where Qdrant uses GPU capabilities.<br>4.24 Basic override file with PostgreSQL deployment and master key automatically generated.<br>4.25 Verifying containers and Helm charts hosted on the SUSE Application Collection<br>4.26 Verifying containers and Helm charts hosted on the SUSE Registry<br>4.27 Discovering attached attestations (optional)<br>4.28 Extracting the CycloneDX SBOM and vulnerability scan (optional)<br>5.1 You already have an OS provisioned on the servers and you want to deploy an RKE2 cluster with a Rancher management server.<br>5.2 You already have an OS provisioned on the...