© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
diff --git a/common/vars.html b/common/vars.html index 58c9e33..5801b16 100644 --- a/common/vars.html +++ b/common/vars.html @@ -120,7 +120,7 @@© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
diff --git a/customizations/customization.html b/customizations/customization.html index 5d7bcea..5985b4a 100644 --- a/customizations/customization.html +++ b/customizations/customization.html @@ -167,7 +167,7 @@© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
diff --git a/customizations/helm.html b/customizations/helm.html index 7e71493..8e22799 100644 --- a/customizations/helm.html +++ b/customizations/helm.html @@ -1318,7 +1318,7 @@© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
diff --git a/device-plugin.html b/device-plugin.html index b849a23..3b08e94 100644 --- a/device-plugin.html +++ b/device-plugin.html @@ -151,7 +151,7 @@© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
diff --git a/genindex.html b/genindex.html index a8cebbe..8159cad 100644 --- a/genindex.html +++ b/genindex.html @@ -125,7 +125,7 @@© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
On this page
Network Operator Deployment with Multiple Resources in RDMA Shared Device Plugin
Network Operator Deployment with a Host Device Network and Macvlan Network
Network Operator Deployment with an IP over InfiniBand (IPoIB) Network
SR-IOV Network Operator Deployment – Parallel Node Configuration for SR-IOV
+Network Operator Deployment with Multiple Resources in RDMA Shared Device Plugin
Network Operator Deployment with a Host Device Network and Macvlan Network
Network Operator Deployment with an IP over InfiniBand (IPoIB) Network
SR-IOV Network Operator Deployment – Parallel Node Configuration for SR-IOV
+SR-IOV Network Operator Deployment – Parallel NIC Configuration for SR-IOV
SR-IOV Network Operator Deployment – SR-IOV Using the systemd Service
Network Operator Deployment with an SR-IOV InfiniBand Network
Network Operator Deployment with an SR-IOV InfiniBand Network with PKey Management
Network Operator Deployment for DPDK Workloads with NicClusterPolicy
SR-IOV Network Operator Deployment – Parallel NIC Configuration for SR-IOV
SR-IOV Network Operator Deployment – SR-IOV Using the systemd Service
Network Operator Deployment with an SR-IOV InfiniBand Network
Network Operator Deployment with an SR-IOV InfiniBand Network with PKey Management
Network Operator Deployment for DPDK Workloads with NicClusterPolicy
Each Network Operator Release has a set of default version values for the various components it deploys. It is recommended that these values will not be changed. Testing and validation were performed with these values, and there is no guarantee of interoperability nor correctness when different versions are used.
-helm repo add nvidia https://helm.ngc.nvidia.com/nvidia
helm repo update
Install Network Operator from the NVIDIA NGC chart using the default values:
helm install network-operator nvidia/network-operator -n nvidia-network-operator --create-namespace --version v24.4.0 --wait-
kubectl -n nvidia-network-operator get pods
Warning
+This feature is supported only for Vanilla Kubernetes deployments with SR-IOV Network Operator.
+Warning
+This mode of operation is not compatible with OFED container.
+Warning
+Tech Preview feature.
+Deploy network-operator by Helm with sriov-network-operator and nv-ipam.
+values.yaml
sriovNetworkOperator:
+ enabled: true
+deployCR: true
+nvIpam:
+ deploy: true
+Enable manageSoftwareBridges featureGate for sriov-network-operator
kubectl patch sriovoperatorconfigs.sriovnetwork.openshift.io -n network-operator default --patch '{ "spec": { "featureGates": { "manageSoftwareBridges": true } } }' --type='merge'
+Create IPPool object for nv-ipam
+apiVersion: nv-ipam.nvidia.com/v1alpha1
+kind: IPPool
+metadata:
+ name: pool1
+ namespace: network-operator
+spec:
+ subnet: 192.168.0.0/16
+ perNodeBlockSize: 100
+ gateway: 192.168.0.1
+ nodeSelector:
+ nodeSelectorTerms:
+ - matchExpressions:
+ - key: node-role.kubernetes.io/worker
+ operator: Exists
+Supported operating systems:
+Ubuntu 22.04
OpenvSwitch from the NVIDIA DOCA for Host package with doca-all or doca-networking profile should be installed on each worker node.
Check NVIDIA DOCA Official installation guide +for details.
+Supported OpenvSwitch dataplanes:
+OVS-kernel
OVS-doca
Check OpenvSwitch Offload document to know about differences.
+These steps are for OVS-kernel data plane, to use OVS-doca follow instructions from the relevant section.
+Configure Open_vSwitch
+ovs-vsctl set Open_vSwitch . other_config:hw-offload=true
+Restart Open_vSwitch
+systemctl restart openvswitch-switch.service
+Create SriovNetworkNodePolicy for selected NIC
+kind: SriovNetworkNodePolicy
+metadata:
+ name: ovs-switchdev
+ namespace: network-operator
+spec:
+ eSwitchMode: switchdev
+ mtu: 1500
+ nicSelector:
+ deviceID: 101d
+ vendor: 15b3
+ nodeSelector:
+ node-role.kubernetes.io/worker: ""
+ numVfs: 4
+ resourceName: switchdev
+ bridge:
+ ovs: {}
+Create OVSNetwork CR
+apiVersion: sriovnetwork.openshift.io/v1
+kind: OVSNetwork
+metadata:
+ name: ovs
+ namespace: network-operator
+spec:
+ networkNamespace: default
+ ipam: |
+ {
+ "type": "nv-ipam",
+ "poolName": "pool1"
+ }
+ resourceName: switchdev
+These steps are for OVS-doca data plane, to use OVS-kernel follow instructions from the relevant section.
+Configure hugepages
+mkdir -p /hugepages
+mount -t hugetlbfs hugetlbfs /hugepages
+echo 4096 > /sys/devices/system/node/node0/hugepages/hugepages-2048kB/nr_hugepages
+Note: for multi CPU system hugepages should be created for each NUMA node: node0, node1, …
+Configure system to create hugepages on boot
+echo "vm.nr_hugepages=8192" > /etc/sysctl.d/99-hugepages.conf
+Note: this example is for a server with two CPU
+Configure Open_vSwitch
+ovs-vsctl --no-wait set Open_vSwitch . other_config:doca-init=true
+ovs-vsctl set Open_vSwitch . other_config:hw-offload=true
+Restart Open_vSwitch
+systemctl restart openvswitch-switch.service
+Create SriovNetworkNodePolicy for selected NIC
+kind: SriovNetworkNodePolicy
+metadata:
+ name: ovs-switchdev
+ namespace: network-operator
+spec:
+ eSwitchMode: switchdev
+ mtu: 1500
+ nicSelector:
+ deviceID: 101d
+ vendor: 15b3
+ nodeSelector:
+ node-role.kubernetes.io/worker: ""
+ numVfs: 4
+ resourceName: switchdev
+ bridge:
+ ovs:
+ bridge:
+ datapathType: netdev
+ uplink:
+ interface:
+ type: dpdk
+Create OVSNetwork CR
+apiVersion: sriovnetwork.openshift.io/v1
+kind: OVSNetwork
+metadata:
+ name: ovs
+ namespace: network-operator
+spec:
+ networkNamespace: default
+ ipam: |
+ {
+ "type": "nv-ipam",
+ "poolName": "pool1"
+ }
+ resourceName: switchdev
+ interfaceType: dpdk
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+ name: ovs-offload
+ labels:
+ app: ovs-offload
+spec:
+ replicas: 2
+ selector:
+ matchLabels:
+ app: ovs-offload
+ template:
+ metadata:
+ labels:
+ app: ovs-offload
+ annotations:
+ k8s.v1.cni.cncf.io/networks: ovs
+ spec:
+ containers:
+ - name: ovs-offload-container
+ command: ["/bin/bash", "-c"]
+ args:
+ - |
+ while true; do sleep 1000; done
+ image: mellanox/rping-test
+ resources:
+ requests:
+ nvidia.com/switchdev: 1
+ limits:
+ nvidia.com/switchdev: 1
+© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
diff --git a/k8s-baremetal-ethernet.html b/k8s-baremetal-ethernet.html index 9f335b5..a9ab6d2 100644 --- a/k8s-baremetal-ethernet.html +++ b/k8s-baremetal-ethernet.html @@ -421,7 +421,7 @@© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
diff --git a/kubernetes-perfomance.html b/kubernetes-perfomance.html index 79951fc..a8c32dc 100644 --- a/kubernetes-perfomance.html +++ b/kubernetes-perfomance.html @@ -154,7 +154,7 @@© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
diff --git a/life-cycle-management.html b/life-cycle-management.html index 26105db..0535c1d 100644 --- a/life-cycle-management.html +++ b/life-cycle-management.html @@ -574,7 +574,7 @@© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
diff --git a/multi-network-pod.html b/multi-network-pod.html index 0d4cfa9..972b8fb 100644 --- a/multi-network-pod.html +++ b/multi-network-pod.html @@ -156,7 +156,7 @@© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
diff --git a/overview.html b/overview.html index 1057bfd..057acfd 100644 --- a/overview.html +++ b/overview.html @@ -140,7 +140,7 @@© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
diff --git a/platform-support.html b/platform-support.html index 639008d..f10092e 100644 --- a/platform-support.html +++ b/platform-support.html @@ -294,7 +294,7 @@© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
diff --git a/release-notes.html b/release-notes.html index 031d314..629c888 100644 --- a/release-notes.html +++ b/release-notes.html @@ -448,7 +448,7 @@© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.
diff --git a/search.html b/search.html index 498f72f..a66e41a 100644 --- a/search.html +++ b/search.html @@ -133,7 +133,7 @@© Copyright 2024, NVIDIA. - Last updated on May 02, 2024. + Last updated on May 21, 2024.