What problem does it solve? Setting up NVIDIA Network Operator networking on Kubernetes requires correctly configuring hardware-specific settings, choosing the right network use case (SR-IOV, InfiniBand, IPoIB, Macvlan, HostDevice), and formatting the error-prone NETOP_NETLIST device list. This Skill guides you through the entire configuration cascade so your global_ops_user.cfg matches your platform and workload. ## Core Features & Use Cases - Platform Config Selection: Copy and customize pre-built configs for DGX (B200, B300, GB200, GB300, H100/H200), Dell PowerEdge, OCI, Lenovo, and SuperMicro platforms. - Use Case Guidance: Choose between sriovnet_rdma, sriovibnet_rdma, hostdev_rdma_sriov, ipoib_rdma_shared_device, and macvlan_rdma_shared_device with correct device formats. - NETOP_NETLIST Formatting: Avoid common mistakes like using PCI BDF addresses for InfiniBand use cases or wrong field counts in device entries. - Feature Flags & Validation: Set OFED_ENABLE, NFD_ENABLE, IPAM_TYPE, and CREATE_CONFIG_ONLY, then generate and review YAML before deploying. - Use Case: You have a DGX GB300 cluster and need SR-IOV InfiniBand RDMA networking. This Skill walks you through copying the DGXGB300 config, selecting sriovibnet_rdma, formatting NETOP_NETLIST with IB interface names, and validating the generated YAML. ## Quick Start Ask the AI to configure netop-tools for a DGX GB300 platform with the sriovibnet_rdma use case and validate the generated network operator YAML.
