k8s 场景异常时的运维处理记录(三)

k8s 场景异常时的运维处理记录（三）

Q1.services中endpoints无法识别

kubectl describe service nginx-v1-svc-deploy

无法通过节点来访问pods 容器服务

正常情况如下：

kubectl describe service nginx-service-nodeport

能绑定到容器里的IP

原因：

yaml文件中的name 标识命名不规范

services文件

nginx-v1-svc-deploy.yaml
apiVersion: v1
kind: Service
metadata:
  name: nginx-v1-svc-deploy 
spec:
  ports:
    - port: 8001
      targetPort: 80
      protocol: TCP
  type: NodePort
  selector:
    name: nginxv1 （原名称：nginx-v1-svc-deploy )

deploy文件

nginx-v1-pods-deploy.yaml
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  name: nginx-v1-pods-deploy
spec:
  replicas: 2
  template:
    metadata:
      labels:
        name: nginxv1（原名称：nginx-v1-svc-deploy )
    spec:
      containers:
        - name: nginxv1（原名称：nginx-v1-svc-deploy )
          image: k8s1:5000/nginx:v1
          ports:
           - containerPort: 80

Q2：master上无法访问cluster IP

因为无路由及kube-proxy模块

Q3.pod 升级回滚错误

[root@k8s1 opt]# kubectl rolling-update nginx-controller nginx-v1 --image=k8s1:5000/nginx:v1 --update-period=10s

error: Found existing in-progress update to image (nginx:v1).

Either continue in-progress update with --image=nginx:v1 or rollback with --rollback

See 'kubectl rolling-update -h' for help and examples.

由于执行过

kubectl rolling-update nginx-controller nginx-v1 --image=nginx:v1

需要回退

kubectl rolling-update nginx-controller nginx-v1 --rollback

Q4.pod 出现 CrashLoopBackOff

查看Pod对应的节点

kubectl get pods -o wide

查看Pod 详细信息

kubectl describe pods nginx-v1-deploy-6d6d7d48c6-vf6xt

提示错误

Back-off restarting failed container

可能升级方式不对使用 editor

更换为：

kubectl set image deployment/nginx-v1-deploy nginx-v1-deploy=k8s4:5000/nginx:v1.12.2

检查进度

kubectl rollout status deployment/nginx-v1-deploy

Q5. 报权限问题

Error from server (Forbidden): Forbidden (user=system:anonymous, verb=get, resource=nodes, subresource=proxy) ( pods/log nginx-v1-deploy-6d6d7d48c6-vf6xt)

临时赋权

kubectl create clusterrolebinding system:anonymous --clusterrole=cluster-admin --user=system:anonymous

Q6：配置kubectl子命令补全

[root@master1 work]# yum install -y bash-completion

[root@master1 work]# source /usr/share/bash-completion/bash_completion

[root@master1 work]# source <(kubectl completion bash)

[root@master1 work]# kubectl completion bash > ~/.kube/completion.bash.inc

[root@master1 work]# source '/root/.kube/completion.bash.inc'

[root@master1 work]# source $HOME/.bash_profile

kubeclt 执行流程