Kubernetes环境下Fedora容器中的MySQL服务无法重启求助

Hey there, let's break down why your existing commands aren't working and walk through the proper fixes:
Most minimal Fedora-based container images skip including systemd or traditional SysV init scripts (the /etc/init.d tools) to keep the image lightweight. Containers are designed to run single processes directly, not managed by a full init system—this is why systemctl is missing and /etc/init.d/mysqld doesn't exist.

Here are the best ways to restart your MySQL service, ordered by Kubernetes-native best practices:

1. Restart the Kubernetes Pod (Recommended) #

Since you're running MySQL in Kubernetes, the most idiomatic approach is to let the orchestration layer handle the restart. Just delete the existing Pod, and your Deployment/StatefulSet will automatically spin up a fresh instance with MySQL running cleanly:

# Replace <mysql-pod-name> with your actual Pod's name
kubectl delete pod <mysql-pod-name>

This ensures the container follows its intended startup process and aligns with how Kubernetes manages workloads.

2. Restart MySQL Directly Inside the Container (If Needed) #

If you need to handle it from within the container shell, use MySQL's native tools or process management:

• Gracefully shut down MySQL first:

  # Enter the container shell if you haven't already
  kubectl exec -it <mysql-pod-name> -- /bin/bash
  
  # Use mysqladmin to trigger a clean shutdown (you'll need your root password)
  mysqladmin -u root -p shutdown
  

• Restart the mysqld process:
  Depending on your container's setup, you can start it directly with:

  mysqld --user=mysql &
  

  Note: If mysqld is the container's main process, shutting it down will likely cause the container to exit, and Kubernetes will restart it automatically anyway—so this is essentially the same as deleting the Pod.

• Alternative: Terminate the mysqld process
  If mysqladmin isn't available, find the process ID and send a graceful termination signal:

  # Locate the mysqld process ID
  ps aux | grep mysqld
  
  # Send a TERM signal to shut it down gracefully (replace <pid> with the actual ID)
  kill -TERM <pid>
  

  Again, this will almost always trigger a container restart via Kubernetes.

内容的提问来源于stack exchange，提问作者Struts