5.5. 配置 SR-IOV Operator
单根 I/O 虚拟化(SR-IOV)通过从单一设备在多个 pod 之间提供共享来提高 NVIDIA GPUDirect RDMA 的性能。
先决条件
- 已安装 SR-IOV Operator。
流程
运行以下命令,验证 Operator 是否已安装并运行
openshift-sriov-network-operator命名空间中的 pod:$ oc get pods -n openshift-sriov-network-operator输出示例
NAME READY STATUS RESTARTS AGE sriov-network-operator-7cb6c49868-89486 1/1 Running 0 22s对于默认的
SriovOperatorConfigCR 以用于 MLNX_OFED 容器,请运行这个命令来更新以下值:apiVersion: sriovnetwork.openshift.io/v1 kind: SriovOperatorConfig metadata: name: default namespace: openshift-sriov-network-operator spec: enableInjector: true enableOperatorWebhook: true logLevel: 2运行以下命令在集群中创建资源:
$ oc create -f sriov-operator-config.yaml输出示例
sriovoperatorconfig.sriovnetwork.openshift.io/default created运行以下命令修补 sriov-operator,以便 MOFED 容器可以使用它:
$ oc patch sriovoperatorconfig default --type=merge -n openshift-sriov-network-operator --patch '{ "spec": { "configDaemonNodeSelector": { "network.nvidia.com/operator.mofed.wait": "false", "node-role.kubernetes.io/worker": "", "feature.node.kubernetes.io/pci-15b3.sriov.capable": "true" } } }'输出示例
sriovoperatorconfig.sriovnetwork.openshift.io/default patched