Docker Auto Restart

This role was implemented to address the classic issue: "Have you tried turning it off and on again?". The problem initially arose with the fetchmail container in Mailu, which fails if only some containers, and not the full docker-compose composition, are restarted.

Overview

This role automates the restart process for all docker-compose instances within a specified directory. It ensures consistent restarts of services while avoiding issues caused by partial restarts.

Features

Automatically detects and restarts docker-compose instances in the given directory.
Uses a Python script to perform service restarts with docker-compose restart.
Integrates with systemd for scheduled or manual execution.
Designed for idempotency and ease of integration.

Requirements

docker-compose must be installed on the target system.
Python 3.x is required to execute the provided script.
This role depends on the system-maintenance-lock role for handling system-wide locking during restarts.

Installation

Clone or include this role in your Ansible project.

Define the required variables in your playbook or inventory:

path_administrator_scripts: "/path/to/administrator/scripts/"
restart_docker_folder: "/path/to/restart/volumes/"
on_calendar_restart_dockers: "daily"

Usage

Include this role in your playbook:

- hosts: all
  roles:
    - docker-auto-restart

The role will:

Set up the necessary directories and scripts.
Configure a systemd service to restart docker-compose instances.
Optionally schedule restarts via a systemd timer.

Acknowledgments

This role was developed with the assistance of ChatGPT, including insights and optimizations from this conversation. Special thanks for providing guidance on error handling, Ansible best practices, and Python integration.

Feel free to contribute or provide feedback via the repository issues page.

2.1 KiB Raw Permalink Blame History