I'm trying to replace the existing default.conf Nginx file with my own one using ConfigMap object but this action leads to "Back-off restarting failed container" error. If I change overwriting directory from conf.d to new nginx.d then the passing my config will be successful. It looks like nginx doesn't allow to rewrite its existing files and directories. Is there any workaround or I made a mistake in my deployment file?

Here are my yaml files. Deployment file

apiVersion: apps/v1
kind: Deployment
metadata:
  name: skadate-deployment
  labels:
    tier: backend
spec:
  replicas: 1
  selector:
    matchLabels:
      app: skadate
      tier: backend
  template:
    metadata:
      labels:
        app: skadate
        tier: backend
    spec:
      volumes:
        - name: skadate-software
          emptyDir: {}
        - name: nginx-config
          configMap:
            name: skadate-test-nginx-config
            items:
              - key: default.conf
                path: default.conf
      containers:
        - name: php
          image: psiloscop/php-hello-world:latest
          ports:
            - containerPort: 81
              name: http
              protocol: TCP
          volumeMounts:
            - name: skadate-software
              mountPath: /var/www/html
          lifecycle:
            postStart:
              exec:
                command: [ "/bin/sh", "-c", "cp -r /app/. /var/www/html" ]
        - name: nginx
          image: nginx:stable
          ports:
            - containerPort: 80
          volumeMounts:
            - name: skadate-software
              mountPath: /var/www/html
            - name: nginx-config
              mountPath: /etc/nginx/conf.d/default.conf
              subPath: default.conf

ConfigMap file

apiVersion: v1
kind: ConfigMap
metadata:
  name: skadate-test-nginx-config
  labels:
    tier: backend
data:
  default.conf : |
    server {
      listen 80;
    
      location / {
        root /var/www/html
        index index.html index.htm;
        try_files $uri $uri/ /index.html;
      }
    }

Here is the Deployment pod logs:

CodePudding user response：

Me too have faced something similar recently. To manage this we can hack the CMD (command) to do the needful.

First create startup script for the nginx process for the nginx container. & this has to be a configmap for sure.

#!/bin/sh
rm -rf /etc/nginx/conf.d/default.conf
cp /tmp/default.conf /etc/nginx/conf.d/default.conf
nginx -g daemon off;

Then use this script for the nginx container in the deployment file. Sometime like below:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: skadate-deployment
  labels:
    tier: backend
spec:
  replicas: 1
  selector:
    matchLabels:
      app: skadate
      tier: backend
  template:
    metadata:
      labels:
        app: skadate
        tier: backend
    spec:
      volumes:
        - name: skadate-software
          emptyDir: {}
        - name: nginx-config
          configMap:
            name: skadate-test-nginx-config
            items:
              - key: default.conf
                path: default.conf
      containers:
        - name: php
          image: psiloscop/php-hello-world:latest
          ports:
            - containerPort: 81
              name: http
              protocol: TCP
          volumeMounts:
            - name: skadate-software
              mountPath: /var/www/html
          lifecycle:
            postStart:
              exec:
                command: [ "/bin/sh", "-c", "cp -r /app/. /var/www/html" ]
        - name: nginx
          image: nginx:stable
          command: ['sh','/script/nginx-start']
          ports:
            - containerPort: 80
          volumeMounts:
            - name: skadate-software
              mountPath: /var/www/html
            - name: nginx-config
              mountPath: /tmp
              subPath: default.conf
            - name: nginx-start-script
              mountpath: /script
              subPath: nginx-start

Here the two important changes are:

1. The new default.conf file we are placing in the /tmp path & will be taken care by the custom startup script to copy into the /etc/nginx/conf.d/default.conf path.
2. A custom script has been added to replace the default.conf file with new one & then start the nginx process

An extra configmap for the custom startup script

apiVersion: v1
kind: ConfigMap
metadata:
  name: nginx-start-script
  labels:
    app: nginx
data:
  nginx-start : |
    #!/bin/sh
    rm -rf /etc/nginx/conf.d/default.conf
    cp /tmp/default.conf /etc/nginx/conf.d/default.conf
    nginx -g daemon off;

Hope this would resolve your issue. Please share the results.