skills.rest
Mellanox
MellanoxOFFICIAL
Explore all skills by this author →
WebsiteGitHub

deploying-network-operator

Deploy NVIDIA Network Operator on Kubernetes clusters via netop-tools scripts.

7|4|Updated Oct 24, 2024
Search Tags:#deployment#kubernetes#helm#rdma#nvidia network operator#sr-iov#netop-tools
One-click install
npx skills add https://github.com/Mellanox/netop-tools --skill deploying-network-operator-mellanox
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: deploying-network-operator
Source: https://github.com/Mellanox/netop-tools/tree/main/skills/deploying-network-operator
Command: npx skills add https://github.com/Mellanox/netop-tools --skill deploying-network-operator-mellanox

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Deploying the NVIDIA Network Operator involves many manual steps: generating Helm values, creating NicClusterPolicy and network CRDs, configuring SR-IOV node pools, and verifying RDMA networking. This Skill guides the full pipeline from configuration generation through Helm install and post-deploy verification, reducing misconfiguration risk. ## Core Features & Use Cases - Config Generation: Runs mk-config.sh to produce values.yaml, NicClusterPolicy.yaml, network CRDs, IPAM pools, and SR-IOV node pool configs from a platform-specific global_ops_user.cfg. - Guided Deployment: Walks through use case selection, dry-run config review (CREATE_CONFIG_ONLY=1), and actual Helm/kubectl deployment (CREATE_CONFIG_ONLY=0). - Troubleshooting & Verification: Provides checks for common failures (missing config, wrong chart version, SR-IOV sync delays, IP pool exhaustion) plus verification commands like getnetwork.sh, checksriovstate.sh, and checkipam.sh. - Use Case: A cluster admin needs RDMA networking for GPU workloads on a DGX B200 cluster. The Skill guides copying the platform config, generating YAML, deploying the operator, and confirming SR-IOV VFs are synced. ## Quick Start Deploy the NVIDIA Network Operator to my Kubernetes cluster using netop-tools with the sriovnet_rdma use case.

Frequently Asked Questions about deploying-network-operator

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I deploy NVIDIA Network Operator on Kubernetes?▼

Source NETOP_ROOT_DIR.sh, copy a platform config to global_ops_user.cfg, run setuc.sh for your use case, generate configs with mk-config.sh, then set CREATE_CONFIG_ONLY=0 and run install/ins-network-operator.sh to deploy via Helm.

How to configure SR-IOV VFs with NVIDIA Network Operator?▼

SR-IOV VF allocation is defined in the sriov-node-pool-config.yaml generated by mk-sriov-node-pool.sh based on your NETOP_NETLIST device list. After deployment, run syncsriov.sh to wait for VF creation, which can take 10 minutes or more.

Why does the deployment create no VFs or apply nothing?▼

The CREATE_CONFIG_ONLY variable defaults to 1, which is a dry run that only generates YAML files. Set CREATE_CONFIG_ONLY=0 before running ins-network-operator.sh to actually execute the Helm install and kubectl apply commands.

Which IPAM should I use, nv-ipam or whereabouts?▼

Use nv-ipam for large clusters and whereabouts for clusters under 60 nodes, controlled by the IPAM_TYPE variable. If IP pools are exhausted, increase NETOP_PERNODE_BLOCKSIZE or expand NETOP_NETWORK_RANGE.

What Network Operator Helm chart versions are supported?▼

netop-tools supports NETOP_VERSION values from 24.7.0 through 26.1.0, with 26.1.0 as the default. A 'Helm chart not found' error usually means the configured version is unavailable.

Related Skills

Explore More Software Engineering Skills→
devendrapratapsinghdevendrapratapsingh
1

provisioning

Automate cloud and bare-metal resource provisioning with Terraform.

Community
Advanced
samuelmukotisamuelmukoti

modal-serverless-gpu

Run ML workloads on serverless GPU cloud with Python-native infrastructure definitions.

Community
Advanced
oracleoracle
785

oke-multihome-deployer

Automate Multus multihome deployment and troubleshooting on OCI Kubernetes Engine.

Official
Advanced
NVIDIANVIDIA
14

k8s-launch-kit-deploy

Apply generated NVIDIA networking manifests to a Kubernetes cluster in dependency order.

Official
Advanced