kubernetes之kube-policy-controller

1.目录

/etc/kubernetes/manifests/kube-policy-controller.json

{
  "kind": "Pod",
  "apiVersion": "v1",
  "metadata": {
    "name": "kube-policy-controller",
    "namespace": "kube-system",
    "creationTimestamp": null,
    "labels": {
      "component": "kube-policy-controller",
      "tier": "control-plane"
    }
  },
  "spec": {
    "volumes": [
      {
        "name": "run-calico",
        "hostPath": {
          "path": "/var/run/calico"
        }
      },
      {
        "name": "log-calico",
        "hostPath": {
          "path": "/var/log/calico"
        }
      },
      {
        "name": "lib-modules",
        "hostPath": {
          "path": "/lib/modules"
        }
      }
    ],
    "containers": [
      {
        "name": "kube-policy-controller",
        "image": "xxx/xxx/kube-policy-controller:v0.5.2-v1",
        "env": [
          {
            "name": "ETCD_ENDPOINTS",
            "value": "http://127.0.0.1:2379"
          },
          {
            "name": "K8S_API",
            "value": "http://127.0.0.1:8080"
          },
          {
            "name": "LEADER_ELECTION",
            "value": "true"
          }
        ],
        "resources": {},
        "volumeMounts": [
          {
            "name": "run-calico",
            "mountPath": "/var/run/calico"
          },
          {
            "name": "log-calico",
            "mountPath": "/var/log/calico"
          },
          {
            "name": "lib-modules",
            "mountPath": "/lib/modules"
          }
        ]
      },
      {
        "name": "leader-elector",
        "image": "xxx/xxx/leader-elector:v0.1.0",
        "args": [
          "--election=calico-policy-election",
          "--election-namespace=kube-system",
          "--http=127.0.0.1:4040"
        ],
        "resources": {},
        "imagePullPolicy": "IfNotPresent",
        "securityContext": {
          "privileged": true
        }
      }
    ],
    "hostNetwork": true
  },
  "status": {}
}

为每一个proxy创建hostendpoint

在k8s集群的控制节点上，为每一个proxy创建hostendpoint，例如为IP地址为192.168.1.110的proxy创建hostEndpoint：

apiVersion: v1
kind: hostEndpoint
metadata:
  name: eth0
  node: 192.168.1.110
  labels:
calico/k8s_ns: kube-system
spec:
  interfaceName: eth0
  expectedIPs:
  - 192.168.1.110
  profiles:
  - k8s_ns.kube-systemk8s

2.集群的控制节点上执行：

calicoctl   create  -f   XXX.yaml

http://cbonte.github.io/haproxy-dconv/1.7/management.html