Ansible启停ssh服务异常：执行停止后sshd仍运行

Let's break down why your SSH service isn't actually stopping when you run that playbook, and fix it step by step.

First: Verify the Correct Service Name #

Different Linux distros use different names for the SSH service:

• Debian/Ubuntu: ssh
• RHEL/CentOS/Fedora: sshd

Run this command on your localhost to confirm the exact name:

systemctl list-unit-files | grep ssh

If it shows sshd.service instead of ssh.service, that's your first fix—update the name field in your Ansible task.

Why Your Original Playbook Fails #

Using with_items to run stop then start back-to-back is the main culprit here. Ansible sends the stop command and immediately sends the start command without waiting for the service to fully shut down. The system might still be processing the stop request when the start command hits, so it looks like the service never stopped.

Fixed Playbook: Separate Tasks with Validation #

Instead of looping, split the tasks and add a check to ensure the service is actually stopped before starting it again. Here's a reliable version:

- hosts: localhost
  become: yes
  tasks:
    - name: Stop SSH service
      service:
        name: ssh  # Replace with sshd if needed
        state: stopped

    - name: Wait for SSH service to fully stop
      command: service ssh status
      register: ssh_stop_status
      failed_when: "'active (running)' in ssh_stop_status.stdout"
      retries: 3
      delay: 2
      until: ssh_stop_status is succeeded

    - name: Start SSH service
      service:
        name: ssh  # Match the name above
        state: started

Alternative: Use the systemd Module (More Reliable for Modern Systems) #

For systems using systemd (most modern distros), the systemd module offers better control and feedback than the generic service module. Try this:

- hosts: localhost
  become: yes
  tasks:
    - name: Stop SSH service
      systemd:
        name: ssh  # Or sshd
        state: stopped
        daemon_reload: yes  # Optional, but useful if configs changed

    - name: Verify SSH service is stopped
      systemd:
        name: ssh
        state: stopped
      register: ssh_systemd_status
      retries: 3
      delay: 2
      until: ssh_systemd_status.status.ActiveState == "inactive"

    - name: Start SSH service
      systemd:
        name: ssh
        state: started

Edge Case: Check if SSH is Configured to Auto-Restart #

In some cases, the SSH service might be set to automatically restart if it stops. Check this with:

systemctl show ssh.service | grep Restart

If you see Restart=always, you'll need to temporarily override this or edit the service file before stopping it. But this is rare for default SSH configurations.

内容的提问来源于stack exchange，提问作者VIVEK GAUR