K8s创建PodSandBox失败：PF设备未找到问题求助

K8s新手，本地搭建2节点minikube集群，从一个PF创建2个VF，希望为Pod配置多网卡并绑定单个VF。已部署SRIOV-CNI、sriov-network-device-plugin（节点已识别到2个可分配VF）及multus组件，但Pod启动失败，事件提示FailedCreatePodSandBox，具体错误为SRIOV-CNI加载配置失败，无法获取VF信息，提示**"PF network device not found"**。

1. minikube集群信息 #

minikube profile list
|----------|-----------|---------|--------------|------|---------|---------|-------|--------|
| Profile  | VM Driver | Runtime |      IP      | Port | Version | Status  | Nodes | Active |
|----------|-----------|---------|--------------|------|---------|---------|-------|--------|
| minikube | docker    | docker  | 192.168.76.2 | 8443 | v1.27.4 | Running |     2 | *      |
|----------|-----------|---------|--------------|------|---------|---------|-------|--------|

2. VF配置信息 #

eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP mode DEFAULT group default qlen 1000
    link/ether 0c:c4:7a:77:f9:80 brd ff:ff:ff:ff:ff:ff
    vf 0     link/ether 1e:4c:7c:6b:7d:0f brd ff:ff:ff:ff:ff:ff, spoof checking on, link-state auto, trust off, query_rss off
    vf 1     link/ether 9e:52:6e:e0:e9:07 brd ff:ff:ff:ff:ff:ff, spoof checking on, link-state auto, trust off, query_rss off

3. SRIOV设备插件ConfigMap配置片段 #

apiVersion: v1
kind: ConfigMap
metadata:
  name: sriovdp-config
  namespace: kube-system
data:
  config.json: |
    {
        "resourceList": [{
                "resourceName": "intel_sriov_netdevice",
                "selectors": {
                    "vendors": ["8086"],
                    "devices": ["15a8"],
                    "drivers": ["ixgbevf"]
                }
            }
    }

4. 组件运行状态 #

SRIOV设备插件Pod运行状态 #

k -n kube-system get pod -l app=sriovdp -o wide
NAME                                   READY   STATUS    RESTARTS   AGE     IP             NODE           NOMINATED NODE   READINESS GATES
kube-sriov-device-plugin-amd64-64bf7   1/1     Running   0          5h41m   192.168.76.2   minikube       <none>           <none>
kube-sriov-device-plugin-amd64-7knxh   1/1     Running   0          5h41m   192.168.76.3   minikube-m02   <none>           <none>

节点可分配VF数量 #

kubectl get node  minikube -o jsonpath='{.status.allocatable}' |jq -r '."intel.com/intel_sriov_netdevice"'
2

Multus组件运行状态 #

k get pod -l app=multus -A -o wide
NAMESPACE     NAME                   READY   STATUS    RESTARTS   AGE     IP             NODE           NOMINATED NODE   READINESS GATES
kube-system   kube-multus-ds-47zg5   1/1     Running   0          5h38m   192.168.76.2   minikube       <none>           <none>
kube-system   kube-multus-ds-rjrzn   1/1     Running   0          5h38m   192.168.76.3   minikube-m02   <none>           <none>

Events:
  Type     Reason                  Age                From               Message
  ----     ------                  ----               ----               -------
  Normal   Scheduled               15s                default-scheduler  Successfully assigned default/testpod1 to minikube-m02
  Normal   AddedInterface          14s                multus             Add eth0 [10.244.1.32/24] from kindnet
  Warning  FailedCreatePodSandBox  14s                kubelet            Failed to create pod sandbox: rpc error: code = Unknown desc = failed to set up sandbox container "0f5d4259dfaa0087fba50ee5c656050fc6af6c01430bdd820e0642fba9d384de" network for pod "testpod1": networkPlugin cni failed to set up pod "testpod1_default" network: plugin type="multus" name="multus-cni-network" failed (add): [default/testpod1/:sriov-network]: error adding container to network "sriov-network": SRIOV-CNI failed to load netconf: LoadConf(): failed to get VF information: "PF network device not found"
  Normal   AddedInterface          13s                multus             Add eth0 [10.244.1.33/24] from kindnet
  Warning  FailedCreatePodSandBox  12s                kubelet            Failed to create pod sandbox: rpc error: code = Unknown desc = failed to set up sandbox container "39c2cc9ceff18eb34fbd4f7b0746fd0807a3999fdcef6f8dd8487b43f812a31a" network for pod "testpod1": networkPlugin cni failed to set up pod "testpod1_default" network: plugin type="multus" name="multus-cni-network" failed (add): [default/testpod1/:sriov-network]: error adding container to network "sriov-network": SRIOV-CNI failed to load netconf: LoadConf(): failed to get VF information: "PF network device not found"
  Normal   AddedInterface          12s                multus             Add eth0 [10.244.1.34/24] from kindnet
  Warning  FailedCreatePodSandBox  11s                kubelet            Failed to create pod sandbox: rpc error: code = Unknown desc = failed to set up sandbox container "20fb94f524db27d3b1bcb0e743c925ddb6af038c9220de28cd84ec92215b0ab3" network for pod "testpod1": networkPlugin cni failed to set up pod "testpod1_default" network: plugin type="multus" name="multus-cni-network" failed (add): [default/testpod1/:sriov-network]: error adding container to network "sriov-network": SRIOV-CNI failed to load netconf: LoadConf(): failed to get VF information: "PF network device not found"

1. 修正SRIOV NetworkAttachmentDefinition配置 #

确保Multus引用的NetworkAttachmentDefinition正确指定PF设备及匹配参数，示例配置如下：

apiVersion: k8s.cni.cncf.io/v1
kind: NetworkAttachmentDefinition
metadata:
  name: sriov-network
spec:
  config: |-
    {
      "cniVersion": "0.3.1",
      "type": "sriov",
      "name": "sriov-network",
      "ipam": {
        "type": "host-local",
        "subnet": "10.56.217.0/24",
        "routes": [
          {
            "dst": "0.0.0.0/0"
          }
        ],
        "gateway": "10.56.217.1"
      },
      "pf": "eth0",
      "vendor": "8086",
      "deviceID": "15a8"
    }

• 重点检查pf字段是否为节点上实际的PF设备名（此处为eth0），vendor和deviceID需与PF的PCI信息一致。

2. 完善SRIOV设备插件ConfigMap #

在ConfigMap的selectors中添加pfNames字段，明确指定PF设备，帮助插件精准识别目标VF：

{
    "resourceList": [{
            "resourceName": "intel_sriov_netdevice",
            "selectors": {
                "vendors": ["8086"],
                "devices": ["15a8"],
                "drivers": ["ixgbevf"],
                "pfNames": ["eth0"]
            }
        }]
}

更新ConfigMap后，重启SRIOV设备插件Pod：

kubectl rollout restart daemonset kube-sriov-device-plugin-amd64 -n kube-system

3. 验证节点PF设备状态 #

在Pod调度的目标节点（minikube-m02）上执行以下命令：

• 确认PF设备存在：ip link show eth0
• 查看PF的PCI信息：lspci -nn | grep Ethernet，验证vendor ID为8086、device ID为15a8
• 确认VF驱动绑定：ethtool -i eth0，PF驱动应为ixgbe（对应VF驱动ixgbevf）

4. 检查SRIOV-CNI权限配置 #

确保SRIOV-CNI的DaemonSet挂载了必要的节点目录（如/sys、/dev），并拥有CAP_NET_ADMIN等权限，典型的DaemonSet权限配置片段：

securityContext:
  capabilities:
    add: ["NET_ADMIN"]
  privileged: true
volumeMounts:
- name: host-sys
  mountPath: /sys
- name: host-dev
  mountPath: /dev
- name: host-etc
  mountPath: /etc/cni/net.d

内容的提问来源于stack exchange，提问作者Verma