kubernetes v1.8.3安装coredns

          coredns的部署yaml文件如下：

apiVersion: v1
kind: ServiceAccount
metadata:
  name: coredns
  namespace: kube-system
  labels:
      kubernetes.io/cluster-service: "true"
      addonmanager.kubernetes.io/mode: Reconcile
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  labels:
    kubernetes.io/bootstrapping: rbac-defaults
    addonmanager.kubernetes.io/mode: Reconcile
  name: system:coredns
rules:
- apiGroups:
  - ""
  resources:
    - endpoints
    - services
    - pods
    - namespaces
  verbs:
    - list
    - watch
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  annotations:
    rbac.authorization.kubernetes.io/autoupdate: "true"
  labels:
    kubernetes.io/bootstrapping: rbac-defaults
    addonmanager.kubernetes.io/mode: EnsureExists
  name: system:coredns
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: system:coredns
subjects:
- kind: ServiceAccount
  name: coredns
  namespace: kube-system
---
apiVersion: v1
kind: ConfigMap
metadata:
  name: coredns
  namespace: kube-system
  labels:
      addonmanager.kubernetes.io/mode: EnsureExists
data:
  Corefile: |
    .:53 {
        errors
        log stdout
        health
        kubernetes cluster.local 172.21.0.0/16
        prometheus
        proxy . /etc/resolv.conf
        cache 30
    }
---
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  name: coredns
  namespace: kube-system
  labels:
    k8s-app: coredns
    kubernetes.io/cluster-service: "true"
    addonmanager.kubernetes.io/mode: Reconcile
    kubernetes.io/name: "CoreDNS"
spec:
  replicas: 1
  selector:
    matchLabels:
      k8s-app: coredns
  template:
    metadata:
      labels:
        k8s-app: coredns
    spec:
      serviceAccountName: coredns
      tolerations:
        - key: node-role.kubernetes.io/master
          effect: NoSchedule
        - key: "CriticalAddonsOnly"
          operator: "Exists"
      containers:
      - name: coredns
        image: registry.cn-hangzhou.aliyuncs.com/pelican/coredns:1.0.1
        imagePullPolicy: IfNotPresent
        resources:
          limits:
            memory: 170Mi
          requests:
            cpu: 100m
            memory: 70Mi
        args: [ "-conf", "/etc/coredns/Corefile" ]
        volumeMounts:
        - name: config-volume
          mountPath: /etc/coredns
        ports:
        - containerPort: 53
          name: dns
          protocol: UDP
        - containerPort: 53
          name: dns-tcp
          protocol: TCP
        - containerPort: 9153
          name: metrics
          protocol: TCP
        livenessProbe:
          httpGet:
            path: /health
            port: 8080
            scheme: HTTP
          initialDelaySeconds: 60
          timeoutSeconds: 5
          successThreshold: 1
          failureThreshold: 5
      dnsPolicy: Default
      volumes:
        - name: config-volume
          configMap:
            name: coredns
            items:
            - key: Corefile
              path: Corefile
---
apiVersion: v1
kind: Service
metadata:
  name: coredns
  namespace: kube-system
  labels:
    k8s-app: coredns
    kubernetes.io/cluster-service: "true"
    addonmanager.kubernetes.io/mode: Reconcile
    kubernetes.io/name: "CoreDNS"
spec:
  selector:
    k8s-app: coredns
  clusterIP: 10.60.0.2
  ports:
  - name: dns
    port: 53
    protocol: UDP
  - name: dns-tcp
    port: 53
    protocol: TCP
  - name: metrics
    port: 9153
    protocol: TCP

其中coredns的configmap参数:

1）errors官方没有明确解释，后面研究

2）log stdout:日志中间件配置为将日志写入STDOUT

3）health:健康检查，提供了指定端口（默认为8080）上的HTTP端点，如果实例是健康的，则返回“OK”。

4）cluster.local：CoreDNS为kubernetes提供的域，10.3.0.0/24这告诉Kubernetes中间件它负责为反向区域提供PTR请求0.0.3.10.in-addr.arpa ..换句话说，这是允许反向DNS解析服务（我们经常使用到得DNS服务器里面有两个区域，即“正向查找区域”和“反向查找区域”，正向查找区域就是我们通常所说的域名解析，反向查找区域即是这里所说的IP反向解析，它的作用就是通过查询IP地址的PTR记录来得到该IP地址指向的域名，当然，要成功得到域名就必需要有该IP地址的PTR记录。PTR记录是邮件交换记录的一种，邮件交换记录中有A记录和PTR记录，A记录解析名字到地址，而PTR记录解析地址到名字。地址是指一个客户端的IP地址，名字是指一个客户的完全合格域名。通过对PTR记录的查询，达到反查的目的。）

5）proxy:这可以配置多个upstream 域名服务器，也可以用于延迟查找 /etc/resolv.conf 中定义的域名服务器

6）cache:这允许缓存两个响应结果，一个是肯定结果（即，查询返回一个结果）和否定结果（查询返回“没有这样的域”），具有单独的高速缓存大小和TTLs。

部署时,写错了configmap的cidr.

通过patch命令修改此configmap

kubectl patch -f coredns.yaml --type="json"  -p='{ "data": { "Corefile": ".:53 {     errors     log stdout     health     kubernetes cluster.local 172.21.0.0/16     prometheus     proxy . /etc/resolv.conf     cache 30 } "} }'

要注意单双引号

kubelet配置此dns的clusterIP地址

测试dns解析.以tiller-deploy的service为例

如上图,tiller-deploy的service名已经正常解析.

参考:

http://www.cnblogs.com/boshen-hzb/p/7511432.html