Docker部署MySQL与phpMyAdmin连接失败，MySQL容器反复重启求助

Hey there, let’s tackle this frustrating issue where your MySQL container keeps restarting and phpMyAdmin can’t connect. That Killed error in your logs is a critical hint—here’s what’s likely going wrong and how to fix it:

Common Causes & Solutions #

1. Out-of-Memory (OOM) Killer is Terminating MySQL #

The most frequent reason for that "Killed" message is that your host machine or container doesn’t have enough memory to run MySQL. The system’s OOM Killer kicks in and terminates the process when it exceeds memory limits.

Fixes:

• Check your host’s memory usage with free -h—if you’re low on available RAM, close other resource-heavy apps or upgrade your host.
• Add memory limits to your MySQL container in docker-compose.yml to prevent it from hogging too much resources:

  services:
    mysql:
      # ... other configs
      deploy:
        resources:
          limits:
            memory: 512M  # Adjust based on your host's capacity
          reservations:
            memory: 256M  # Guarantees this much memory for the container
  

2. Permission Issues on Mounted Data Volumes #

If you’re mounting a local directory as a MySQL data volume, incorrect file permissions can prevent MySQL from writing to its data directory, leading to a crash.

Fixes:

• On your host machine, set the correct ownership for the data directory (MySQL uses UID 999 by default):

  chown -R 999:999 ./mysql-data
  

• If you don’t want to modify host permissions, you can temporarily run the container as root for testing (not recommended for production):

  services:
    mysql:
      # ... other configs
      user: "root"
  

3. Conflicting Startup Parameters or Outdated Images #

Your log mentions --daemonize --skip-networking—some newer MySQL official images handle background execution internally, so adding these parameters can cause conflicts. Using an untested or outdated image might also lead to unexpected crashes.

Fixes:

• Use a stable, official MySQL image (e.g., mysql:8.0) instead of generic or outdated ones.
• Avoid modifying the default entrypoint parameters unless you’re sure you need to. The official image’s entrypoint already manages MySQL’s startup correctly.

Example Working docker-compose.yml #

Here’s a polished version of your compose file that addresses the above issues:

version: '3.3'
services:
  mysql:
    image: mysql:8.0
    container_name: mysql_db
    restart: unless-stopped
    environment:
      MYSQL_ROOT_PASSWORD: your_secure_root_password
      MYSQL_DATABASE: your_default_db
      MYSQL_USER: your_db_user
      MYSQL_PASSWORD: your_user_password
    volumes:
      - ./mysql-data:/var/lib/mysql
    ports:
      - "3306:3306"
    deploy:
      resources:
        limits:
          memory: 512M
        reservations:
          memory: 256M

  phpmyadmin:
    image: phpmyadmin/phpmyadmin
    container_name: phpmyadmin
    restart: unless-stopped
    environment:
      PMA_HOST: mysql  # Uses the MySQL service name for internal Docker DNS
      PMA_PORT: 3306
      PMA_USER: root
      PMA_PASSWORD: your_secure_root_password
    ports:
      - "8080:80"
    depends_on:
      - mysql

How to Test the Fix #

1. Stop and remove existing containers (note: -v deletes volumes—backup data first if needed):

   docker-compose down -v
   

2. Start the containers again:

   docker-compose up -d
   

3. Check the MySQL container status and logs:

   docker ps  # Should show "Up" status for MySQL
   docker-compose logs mysql  # Verify no more "Killed" errors
   

Once MySQL is running, phpMyAdmin should be able to connect using the service name mysql as the host.

内容的提问来源于stack exchange，提问作者aeemk