2021-10-21

要把一个h5的文件夹做成个镜像然后部署

1. 首先整个空文件夹，把h5文件放进去，写个dockerfile

   mkdir test-h5
   cd test-h5/
   vim Dockerfile
     FROM nginx:1.18.0-alpine
     RUN mkdir -p /mnt/nginx/html
     COPY ./h5 /mnt/nginx/html/h5              #./h5是要copy的文件夹  /mnt/nginx/html/h5复制文件夹到文件夹
     ENV TZ=Asia/Shanghai
     CMD ["nginx","-g","daemon off;"]          #nginx前台运行，CMD后面有空格！！！
     
     #具体还有啥命令自己百度
    在dockerfile所在的文件夹下，生成镜像
     docker build -t test-h5:v1 .
   

2. 开始写deployment.yaml,service.yaml

   （参考：https://blog.csdn.net/dy1316434466/article/details/105440172）

   vim test-h5-deployment.yaml
   apiVersion: apps/v1
   kind: Deployment
   metadata:
     name: test-h5
     namespace: default
   spec:
     replicas: 1
     selector:
       matchLabels:
         app: test-h5
         group: default-app
     template:
       metadata:
         labels:
           app: test-h5
           group: default-app
       spec:
         containers:
         - name: web   
           image: test-h5:v1  #镜像
           imagePullPolicy: Always 
           resources:
             limits:
               cpu: 300m
               memory: 100Mi
             requests:
               cpu: 100m
               memory: 50Mi
           workingDir: /mnt/nginx/html
           ports:
           - name: http-port
             containerPort: 80
           livenessProbe:
             tcpSocket:
               port: http-port
             initialDelaySeconds: 5
             periodSeconds: 30
             failureThreshold: 5
           readinessProbe:
             tcpSocket:
               port: http-port
             initialDelaySeconds: 2
             periodSeconds: 5
             successThreshold: 2
             failureThreshold: 5
           volumeMounts:
           - name: test-h5
             mountPath: /etc/nginx/nginx.conf
             subPath: nginx.conf
             readOnly: false
           - name: test-h5
             mountPath: /etc/nginx/conf.d/default.conf
             subPath: test-h5.conf
             readOnly: false
           - name: log-path
             mountPath: /var/log/nginx
             readOnly: false
         hostname: test-h5
         imagePullSecrets:
         - name: [secret-name]   #自己的secret-name
         restartPolicy: Always
         volumes:
         - name: test-h5
           configMap:
             name: test-h5
             items:
             - key: nginx.conf
               path: nginx.conf
             - key: test-h5.conf
               path: test-h5.conf
         - name: log-path
           hostPath:
             path: /home/logs/test-h5
             type: DirectoryOrCreate
   
   

   vim test-h5-service.yaml
   apiVersion: v1
   kind: Service
   metadata:
     name: test-h5
     namespace: default
   spec:
     type: NodePort
     selector:
       app: test-h5
     ports:
     - port: 80
       nodePort: 30099
       name: http-port
       targetPort: http-port
   
   

3. 部署

    kubectl create -f ./test-h5-deployment.yaml -f ./test-h5-service.yaml
   

4. 这个时候通过kubectl get pod 发现是在等待容器的创建

   docker images 发现镜像是存在的，那么就考虑deployment.yaml了。镜像是FROM nginx:1.18.0-alpine，需要替换nginx.conf

   vim test-h5-configmap.yaml
   apiVersion: v1
   kind: ConfigMap
   metadata:
     name: test-h5
     namespace: default
   data:
     nginx.conf: |    
       user  root;
       worker_processes  1;
       
       error_log  /var/log/nginx/error.log warn;
       pid        /var/run/nginx.pid;    
       
       events {
           worker_connections  5000;
       }
       
       http {
           include       /etc/nginx/mime.types;
           default_type  application/octet-stream;
           
           log_format  main  '$remote_addr - $remote_user [$time_local] "$request" '
                             '$status $body_bytes_sent "$http_referer" '
                             '"$http_user_agent" "$http_x_forwarded_for"';
           
           access_log  /var/log/nginx/access.log  main;
           
           sendfile        on;
           #tcp_nopush     on;
           
           keepalive_timeout  65;
           
           gzip  on;
           
           include /etc/nginx/conf.d/*.conf;
           
        }  
     test-h5.conf: |
       server {
           listen     80;
           server_name  _;
   
           location / {
               try_files $uri $uri/ /;
               root        /mnt/nginx/html/h5;
               index index.html;
               add_header Last-Modified $date_gmt;
               add_header Cache-Control 'no-store, no-cache,must-revalidate, proxy-revalidate, max-age=0';
               if_modified_since off;
               expires off;
               etag off;                
           }
           location ~* ^.+\.(jpg|jpeg|gif|png|ico|css|js)$ {
               root        /mnt/nginx/html/h5;
               access_log  off;
               expires 1d; 
           }
           error_page  404              /404.html;
                   location = /404-light.html {
           }
       }    
   
   

5. apply一下

   kubectl apply -f ./test-h5-configmap.yaml
   

6. 这个时候发现镜像拉取失败，但是docker images中有，也不知道为啥，那就把它推远程吧
   （似乎找到原因了：imagePullSecrets: restartPolicy: Always。Always：是拉取远程仓库的镜像，如果远程仓库不存在，pod会报错；IfNotPresent：优先使用本地镜像，本地不存在回去远程仓库拉；Never：只使用本地镜像）

   docker images     #查看生成的镜像所对应的ImageId
   docker tag [ImageId] registry.cn-hangzhou.aliyuncs.com/***/test-h5:v1
   docker push registry.cn-hangzhou.aliyuncs.com/***/test-h5:v1
   #(要记得把deployment中的镜像名称改成这个)
   

关于namespace的问题，如果在deployment中配置了namespace那么在service和configmap都需要，这个决定这个服务跑在哪个命名空间下。

关于h5文件更新问题：这个镜像是根据文件夹来的，文件夹每改一下，镜像就要更改一下。docker build -t test-h5:v1 .这条命令生成test-h5镜像能够覆盖原有的，再进行推送，推送的镜像会覆盖原有的镜像。然后把部署的服务redeploy一下。