Exécutez n'importe quel Skill dans Manus
en un clic

Exécutez n'importe quel Skill dans Manus en un clic

configuring-netop-platform

Étoiles7

Forks4

Mis à jour9 avril 2026 à 15:23

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).

Installation

Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.

Exécuter dans Manus

Source

Mellanox

Mellanox/netop-tools

Ouvrir le dépôt GitHub Voir les dépôts du créateur

Téléchargement

Exécuter dans Manus

Métiers associésSOC

Basé sur la classification professionnelle SOC

Administrateurs de réseaux et de systèmes informatiquesProfessions informatiques et mathématiques·SOC 15-1244

SKILL.md

readonly

name	configuring-netop-platform
description	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).

Configuring netop-tools Platform

Set up configuration for your hardware platform and networking use case.

Configuration Cascade

Priority (highest to lowest): ENV vars > global_ops_user.cfg > usecase/{USECASE}/netop.cfg > global_ops.cfg defaults

Step 1: Copy Platform Config

source NETOP_ROOT_DIR.sh

# Available platforms:
ls config/
# bcm11  dell  dgx  examples  igx  kvm  lenovo  oci  pdx  smc

# Copy for your hardware:
cp config/dgx/global_ops_user.cfg.DGXGB300.bcm global_ops_user.cfg
# Edit global_ops_user.cfg

Platform Quick Reference

Platform	Key Config	NIC	Typical Use Case
DGX B200/B300	`config/dgx/*.bcm`	ConnectX-7/8	`sriovibnet_rdma` or `sriovnet_rdma`
DGX GB200	`config/dgx/*.DGXGB200.bcm`	ConnectX-7, 16 VFs	`sriovibnet_rdma`
DGX GB300	`config/dgx/*.DGXGB300.bcm`	ConnectX-8, 16 VFs	`sriovibnet_rdma`
DGX H100/H200	`config/dgx/*.bcm`	ConnectX-7	`sriovnet_rdma`
Dell PowerEdge	`config/dell/*`	ConnectX-7	`sriovnet_rdma`
OCI	`config/oci/*`	ConnectX, 16 VFs	`sriovnet_rdma`
Examples	`config/examples/*`	Generic	One per use case

Step 2: Choose Use Case

Use Case	Network Type	VFs	Device Format	When to Use
`sriovnet_rdma`	SR-IOV Ethernet	8	PCI BDF: `0000:08:00.0`	Most common, RoCE networks
`sriovibnet_rdma`	SR-IOV InfiniBand	8	IB interface: `ibs0f1`	InfiniBand fabrics
`hostdev_rdma_sriov`	HostDevice	8	Multi-PCI: `0000:07:00.0,0000:08:00.0`	Exclusive device access
`ipoib_rdma_shared_device`	IPoIB shared	0	IB: `ibs0f0` (field3=HCAMAX)	Shared IB, no VFs
`macvlan_rdma_shared_device`	Macvlan shared	0	Eth: `ens2f0np0` (field3=HCAMAX)	Shared Ethernet, no VFs

Step 3: Configure NETOP_NETLIST

This is the most error-prone setting. Format: ( index,,,device_id )

# SR-IOV Ethernet — PCI BDF addresses
NETOP_NETLIST=( a,,,0000:08:00.0 b,,,0000:86:00.1 )

# SR-IOV InfiniBand — IB interface names
NETOP_NETLIST=( a,,,ibs0f1 b,,,ibs1f1 )

# HostDevice — multiple PCI per entry
NETOP_NETLIST=( a,,,0000:07:00.0,0000:08:00.0 )

# Shared device (field3 = HCAMAX, typically 63)
NETOP_NETLIST=( a,,63,ibs0f0 b,,63,ibs0f1 )

Common mistakes: Using PCI BDF for IB use cases, missing commas, wrong field count.

Step 4: Set Feature Flags

Variable	Default	Set When
`OFED_ENABLE`	`true`	`false` if using kernel OFED (not container)
`NFD_ENABLE`	`true`	`false` if GPU-operator runs NFD
`NIC_CONFIG_ENABLE`	`false`	`true` for firmware tuning (GB300, etc.)
`IPAM_TYPE`	`nv-ipam`	`whereabouts` for <60 nodes, `dhcp` for external DHCP
`NETOP_BCM_CONFIG`	`false`	`true` for multi-device combined YAML (DGX)
`CREATE_CONFIG_ONLY`	`1`	`0` to actually deploy

Step 5: Validate

source global_ops.cfg
# Generate config without deploying:
./install/ins-network-operator.sh
# Review generated YAML:
ls -la usecase/${USECASE}/*.yaml

# Python CLI validation:
python3 python_tools/netop_tools.py config validate
python3 python_tools/netop_tools.py config show

Plus depuis ce dépôt

même dépôt

deploying-network-operator

Mellanox/netop-tools

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.

2026-04-097

managing-netop-devices

Mellanox/netop-tools

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.

2026-04-097

testing-netop-configs

Mellanox/netop-tools

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.

2026-04-097

troubleshooting-network-operator

Mellanox/netop-tools

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.

2026-04-097

upgrading-network-operator

Mellanox/netop-tools

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.

2026-04-097

name	configuring-netop-platform
description	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).

Configuring netop-tools Platform

Set up configuration for your hardware platform and networking use case.

Configuration Cascade

Priority (highest to lowest): ENV vars > global_ops_user.cfg > usecase/{USECASE}/netop.cfg > global_ops.cfg defaults

Step 1: Copy Platform Config

source NETOP_ROOT_DIR.sh

# Available platforms:
ls config/
# bcm11  dell  dgx  examples  igx  kvm  lenovo  oci  pdx  smc

# Copy for your hardware:
cp config/dgx/global_ops_user.cfg.DGXGB300.bcm global_ops_user.cfg
# Edit global_ops_user.cfg

Platform Quick Reference

Platform	Key Config	NIC	Typical Use Case
DGX B200/B300	`config/dgx/*.bcm`	ConnectX-7/8	`sriovibnet_rdma` or `sriovnet_rdma`
DGX GB200	`config/dgx/*.DGXGB200.bcm`	ConnectX-7, 16 VFs	`sriovibnet_rdma`
DGX GB300	`config/dgx/*.DGXGB300.bcm`	ConnectX-8, 16 VFs	`sriovibnet_rdma`
DGX H100/H200	`config/dgx/*.bcm`	ConnectX-7	`sriovnet_rdma`
Dell PowerEdge	`config/dell/*`	ConnectX-7	`sriovnet_rdma`
OCI	`config/oci/*`	ConnectX, 16 VFs	`sriovnet_rdma`
Examples	`config/examples/*`	Generic	One per use case

Step 2: Choose Use Case

Use Case	Network Type	VFs	Device Format	When to Use
`sriovnet_rdma`	SR-IOV Ethernet	8	PCI BDF: `0000:08:00.0`	Most common, RoCE networks
`sriovibnet_rdma`	SR-IOV InfiniBand	8	IB interface: `ibs0f1`	InfiniBand fabrics
`hostdev_rdma_sriov`	HostDevice	8	Multi-PCI: `0000:07:00.0,0000:08:00.0`	Exclusive device access
`ipoib_rdma_shared_device`	IPoIB shared	0	IB: `ibs0f0` (field3=HCAMAX)	Shared IB, no VFs
`macvlan_rdma_shared_device`	Macvlan shared	0	Eth: `ens2f0np0` (field3=HCAMAX)	Shared Ethernet, no VFs

Step 3: Configure NETOP_NETLIST

This is the most error-prone setting. Format: ( index,,,device_id )

# SR-IOV Ethernet — PCI BDF addresses
NETOP_NETLIST=( a,,,0000:08:00.0 b,,,0000:86:00.1 )

# SR-IOV InfiniBand — IB interface names
NETOP_NETLIST=( a,,,ibs0f1 b,,,ibs1f1 )

# HostDevice — multiple PCI per entry
NETOP_NETLIST=( a,,,0000:07:00.0,0000:08:00.0 )

# Shared device (field3 = HCAMAX, typically 63)
NETOP_NETLIST=( a,,63,ibs0f0 b,,63,ibs0f1 )

Common mistakes: Using PCI BDF for IB use cases, missing commas, wrong field count.

Step 4: Set Feature Flags

Variable	Default	Set When
`OFED_ENABLE`	`true`	`false` if using kernel OFED (not container)
`NFD_ENABLE`	`true`	`false` if GPU-operator runs NFD
`NIC_CONFIG_ENABLE`	`false`	`true` for firmware tuning (GB300, etc.)
`IPAM_TYPE`	`nv-ipam`	`whereabouts` for <60 nodes, `dhcp` for external DHCP
`NETOP_BCM_CONFIG`	`false`	`true` for multi-device combined YAML (DGX)
`CREATE_CONFIG_ONLY`	`1`	`0` to actually deploy

Step 5: Validate

source global_ops.cfg
# Generate config without deploying:
./install/ins-network-operator.sh
# Review generated YAML:
ls -la usecase/${USECASE}/*.yaml

# Python CLI validation:
python3 python_tools/netop_tools.py config validate
python3 python_tools/netop_tools.py config show