菜单
Deploy NVIDIA Network Operator to a Kubernetes cluster via netop-tools. Use when installing network operator, setting up RDMA networking, deploying SR-IOV, running ins-network-operator.sh, generating Helm values, applying network CRDs, or troubleshooting deployment failures.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Configure netop-tools for a specific hardware platform and use case. Use when setting up global_ops_user.cfg, choosing a use case (sriovnet_rdma, sriovibnet_rdma, hostdev, ipoib, macvlan), selecting IPAM type, configuring NETOP_NETLIST device list, setting feature flags, or switching between platforms (DGX, Dell, Lenovo, OCI, SuperMicro).
Manage SR-IOV virtual functions, PCI devices, link speed, and RDMA testing for NVIDIA Network Operator. Use when setting VFs, querying PCI devices, forcing link speed, running RDMA bandwidth tests, checking RDMA capability, managing node labels and taints, or cordoning nodes.
Run and create tests for netop-tools configuration generation. Use when running unitest.sh, validating YAML generation, creating new test cases, checking test baselines, debugging test failures, or verifying config changes don't break existing tests.
Diagnose and fix NVIDIA Network Operator issues in Kubernetes. Use when pods can't get RDMA network, SR-IOV VFs not showing, IP allocation failures, network attachment missing, must-gather diagnostics, operator pods crashing, NicClusterPolicy errors, or IPAM pool exhaustion.
Upgrade NVIDIA Network Operator to a new version. Use when changing NETOP_VERSION, running upgrade-network-operator.sh, migrating between operator versions, handling CRD schema changes, or rolling back a failed upgrade.
| name | deploying-network-operator |
| description | Deploy NVIDIA Network Operator to a Kubernetes cluster via netop-tools. Use when installing network operator, setting up RDMA networking, deploying SR-IOV, running ins-network-operator.sh, generating Helm values, applying network CRDs, or troubleshooting deployment failures. |
Guide the full deployment pipeline from config generation through Helm install and CRD application.
source NETOP_ROOT_DIR.sh
# global_ops_user.cfg must exist and be edited for the target platform
source global_ops.cfg
# 1. Set use case
./setuc.sh ${USECASE}
# 2. Generate all config (dry run by default: CREATE_CONFIG_ONLY=1)
cd usecase/${USECASE}
${NETOP_ROOT_DIR}/ops/mk-config.sh
# 3. Review generated YAML before deploying
ls -la *.yaml
# 4. Deploy (set CREATE_CONFIG_ONLY=0 to actually run helm/kubectl)
export CREATE_CONFIG_ONLY=0
${NETOP_ROOT_DIR}/install/ins-network-operator.sh
| Script | Output | Purpose |
|---|---|---|
ops/mk-values.sh | values.yaml | Helm values (feature flags, images) |
ops/mk-nic-cluster-policy.sh | NicClusterPolicy.yaml | NicClusterPolicy CRD |
ops/mk-network-cr.sh | network.yaml + ippool-*.yaml | Network + IPAM CRDs per device |
ops/mk-sriov-node-pool.sh | sriov-node-pool-config.yaml | SR-IOV VF allocation |
ops/mk-nic-config.sh | nic-config-crd-{type}.yaml | NIC firmware config (if NIC_CONFIG_ENABLE=true) |
python3 python_tools/netop_tools.py install helm
python3 python_tools/netop_tools.py install chart
python3 python_tools/netop_tools.py install network-operator
python3 python_tools/netop_tools.py install calico
python3 python_tools/netop_tools.py install crds
kubectl get pods -n ${NETOP_NAMESPACE}
./ops/getnetwork.sh
./ops/checksriovstate.sh
./ops/syncsriov.sh # Wait for SR-IOV sync (can take 10+ min)
./ops/checkipam.sh
| Symptom | Cause | Fix |
|---|---|---|
| "User configuration file not found" | Missing global_ops_user.cfg | Copy from config/{platform}/ |
| Helm chart not found | Wrong NETOP_VERSION | Check available versions: 24.7.0 through 26.1.0 |
| No VFs created | CREATE_CONFIG_ONLY=1 (default) | Set CREATE_CONFIG_ONLY=0 before deploy |
| SR-IOV sync stuck | Normal — can take 10 min | Run ./ops/syncsriov.sh to wait |
| Pod network not attached | Wrong namespace | Ensure pod and network CRD are in same namespace |
| IP pool exhausted | NETOP_PERNODE_BLOCKSIZE too small | Increase blocksize or NETOP_NETWORK_RANGE |
CREATE_CONFIG_ONLY — 1 (dry run) or 0 (deploy). Default is 1.NETOP_VERSION — Helm chart version (default 26.1.0)USECASE — Use case name (default sriovnet_rdma)NETOP_NETLIST — Device list (critical, format varies by use case)IPAM_TYPE — nv-ipam (large clusters) or whereabouts (<60 nodes)