> This page is for Self-Hosted, version v3.0 (default). > For other versions, use one of these documentation indexes: > - v3.0 (default): https://docs.nvidia.com/run-ai/self-hosted/v3.0/llms.txt > - v2.26: https://docs.nvidia.com/run-ai/self-hosted/v2.26/llms.txt > - v2.25: https://docs.nvidia.com/run-ai/self-hosted/v2.25/llms.txt > - v2.24: https://docs.nvidia.com/run-ai/self-hosted/v2.24/llms.txt > For clean Markdown of any page, append .md to the page URL. > For a complete documentation index, see https://docs.nvidia.com/run-ai/llms.txt. > For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/run-ai/_mcp/server. # Gov-Ready Installation A Gov-ready installation follows the standard NVIDIA Run:ai installation procedures. You can install using [Guided Installation](/run-ai/self-hosted/documentation/installation-and-configuration/install/guided-installation) (recommended) or [Manual Helm](/run-ai/self-hosted/documentation/installation-and-configuration/install/manual-helm/preparations), on Kubernetes or OpenShift, in connected or air-gapped environments. The standard system requirements, control plane and cluster installation flow, and post-installation steps all apply, with the Gov-ready additions described on this page. Gov-ready adds the following requirements: - **Third-party services** - Use FIPS-validated hardened images and an external PostgreSQL instance. - **GPU Operator and Network Operator** - Install the Gov-ready versions of both operators according to their respective documentation, [GPU Operator](https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/latest/install-gpu-operator-gov-ready.html) and [Network Operator](https://docs.nvidia.com/networking/display/kubernetes2670/install-network-operator-gov-ready.html). - **FIPS mode** - Enable FIPS mode when installing the control plane and the cluster. Each installation procedure identifies where Gov-ready-specific configuration is required, so follow the procedure for your installation path and apply the Gov-ready steps where indicated. For an overview of Gov-ready and what it provides, refer to [Gov-Ready](/run-ai/self-hosted/documentation/installation-and-configuration/overview#gov-ready). ## System Requirements Review the [system requirements](/run-ai/self-hosted/documentation/installation-and-configuration/system-requirements) for your deployment model, infrastructure, and networking setup before you begin. ## Third-Party Services Gov-ready uses an external PostgreSQL instance and FIPS-validated hardened images for the third-party services. The hardened images are Docker Hardened Images (DHI), which you pull from Docker Hub (`docker.io`) from repositories in your Docker Hub organization, using your Docker Hub credentials. For the image registry settings, registry credentials, and Helm values each service requires, refer to [Gov-Ready Third-Party Services](/run-ai/self-hosted/documentation/installation-and-configuration/system-requirements/third-party-services/gov-ready-third-party-services). Pass the Gov-ready values to the installation as custom Helm values: - **Guided Installation** - Add the values under `controlPlane.customHelmValues` and `cluster.customHelmValues` in the installer configuration. - **Manual Helm** - Save the values in a values file and pass the file with `-f` when installing the control plane and cluster. In air-gapped environments, pass it in addition to the values files that `airgap load` generates. The hardened images are provided differently in connected and air-gapped environments: - **Connected** - Create the registry secret for the hardened images before installation, and set the image values. - **Air-gapped** - Add the hardened images and your Docker Hub credentials to the pack configuration when preparing the air-gapped bundle. You do not create a Kubernetes secret or set the image values. The load step generates them. Set only the other values, such as the external PostgreSQL connection. ## Prerequisites and Features Install the prerequisites and features as described in the installation procedure, except for the GPU Operator and Network Operator. Both provide Gov-ready versions, so install each according to its Gov-ready documentation, including any required preparation and Helm values. Refer to the Gov-ready documentation of the [NVIDIA GPU Operator](https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/latest/install-gpu-operator-gov-ready.html) and the [NVIDIA Network Operator](https://docs.nvidia.com/networking/display/kubernetes2670/install-network-operator-gov-ready.html). - **Guided Installation** - Pass the required operator values under the operator's `customHelmValues`. Refer to [Prerequisites and Features](/run-ai/self-hosted/documentation/installation-and-configuration/install/guided-installation/prerequisites-and-features). In air-gapped environments, install the operators before running the installer. - **Manual Helm** - Install the operators with the required values. Refer to [Prerequisites and Features](/run-ai/self-hosted/documentation/installation-and-configuration/install/manual-helm/prerequisites-and-features). ## Install the Control Plane and the Cluster In addition to the third-party service values, enable FIPS mode when installing both the control plane and the cluster by setting `global.fipsMode: only` in the control plane and cluster configuration. - **Guided Installation** - Add `global.fipsMode: only` under both `controlPlane.customHelmValues` and `cluster.customHelmValues` in the installer configuration. - **Manual Helm** - Add `--set global.fipsMode=only` to the `helm upgrade` commands for the control plane and the cluster. ## Next Steps Continue with the installation procedure for your selected installation method: - [Guided Installation](/run-ai/self-hosted/documentation/installation-and-configuration/install/guided-installation) - Recommended for most installations. - [Manual Helm](/run-ai/self-hosted/documentation/installation-and-configuration/install/manual-helm/preparations) - For organizations that manage releases through their own tooling.