GPU Operator on OpenShift
- Introduction
- Prerequisites
- Overview
- Enabling a Cluster-wide entitlement
- Obtaining an entitlement certificate
- Installing the Node Feature Discovery (NFD) Operator
- Installing the NVIDIA GPU Operator
- MIG Support in OpenShift Container Platform
- Cleanup
- Troubleshooting
- Deploy GPU Operators in a disconnected or airgapped environment
- Introduction
- Prerequisites
- Set up a basic HTTP Server
- Check the version of RHEL being used in the cluster
- Mirror the RPM packages
- Creating a private registry
- Authenticate the mirror registry
- Configuring credentials that allow images to be mirrored
- Mirror the Operator catalogs on a disconnected cluster
- Disabling the default OperatorHub sources
- Pruning an index image
- Mirror Node Feature Discovery and the NVIDIA GPU Operator Catalog
- Creating a catalog from an index image
- Verify the mirrored catalog source
- Deploy the Node Feature Discovery Operator
- Deploy the NVIDIA GPU Operator