Optimized healing function for port conflicts

roles/heal-docker/README.md

@@ -1,2 +1,28 @@
-# heal-docker
-docker-compose restart for containers which are unhealty or excited
+# Docker Healer 🩺
+
+## Description
+
+This Ansible role automatically restarts Docker Compose configurations with exited or unhealthy containers on Arch Linux systems. It ensures the stability of containerized workloads by recovering from common error conditions like port binding issues.  
+
+## Overview
+
+Tailored for Arch Linux, this role monitors containers for failure states and initiates a controlled restart of affected Compose configurations. If port conflicts prevent recovery, the role stops the affected stack, restarts Docker, and recreates the container environment.
+
+## Purpose
+
+The purpose of this role is to provide automated healing for Docker Compose environments, minimizing manual recovery effort and reducing downtime.
+
+## Features
+
+- **Container Health Monitoring:** Detects unhealthy or exited containers.
+- **Automated Recovery:** Restarts failed containers and resolves port binding issues.
+- **Run-once Setup Logic:** Ensures idempotent execution by controlling task flow with internal flags.
+- **System Role Integration:** Seamlessly integrates with CyMaIS system maintenance logic.
+
+## Credits 📝
+
+Developed and maintained by **Kevin Veen-Birkenbach**.  
+Learn more at [www.veen.world](https://www.veen.world)
+
+Part of the [CyMaIS Project](https://github.com/kevinveenbirkenbach/cymais)  
+License: [CyMaIS NonCommercial License (CNCL)](https://s.veen.world/cncl)

roles/heal-docker/files/heal-docker.py

@@ -12,10 +12,11 @@ def bash(command):
     process = subprocess.Popen([command], stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=True)
     out, err = process.communicate()
     stdout = out.splitlines()
+    stderr = err.decode("utf-8").strip()  # decode stderr
     output = [line.decode("utf-8") for line in stdout]
-    if process.wait() > bool(0):
+    if process.returncode > 0:
         print(command, out, err)
-        raise Exception("Exitcode is greater than 0")
+        raise Exception(stderr)  # pass the actual error text
     return output
 
 def list_to_string(lst):
@@ -60,10 +61,22 @@ def main(base_directory):
         
         if compose_file_path:
             print("Restarting unhealthy container in:", compose_file_path)
-            print_bash(f'cd {os.path.dirname(compose_file_path)} && docker-compose -p "{repo}" restart')
+            project_path = os.path.dirname(compose_file_path)
+            try:
+                print_bash(f'cd {project_path} && docker-compose -p "{repo}" restart')
+            except Exception as e:
+                if "port is already allocated" in str(e):
+                    print("Detected port allocation problem. Executing recovery steps...")
+                    print_bash(f'cd {project_path} && docker-compose down')
+                    print_bash('systemctl restart docker')
+                    print_bash(f'cd {project_path} && docker-compose -p "{repo}" up -d')
+                else:
+                    print("Unhandled exception during restart:", e)
+                    errors += 1
         else:
             print("Error: Docker Compose file not found for:", repo)
             errors += 1
+
     
     print("Finished restart procedure.")
     exit(errors)

roles/heal-docker/handlers/main.yml

@@ -1,4 +1,5 @@
-- name: "reload heal-docker.cymais.service"
+- name: restart heal-docker.cymais.service
   systemd:
     name: heal-docker.cymais.service
+    state: restarted
     daemon_reload: yes

roles/heal-docker/meta/main.yml

@@ -1,2 +1,26 @@
+---
+galaxy_info:
+  author: "Kevin Veen-Birkenbach"
+  description: "Automated recovery for unhealthy or exited Docker Compose containers."
+  license: "CyMaIS NonCommercial License (CNCL)"
+  license_url: "https://s.veen.world/cncl"
+  company: | 
+    Kevin Veen-Birkenbach
+    Consulting & Coaching Solutions
+    https://www.veen.world
+  min_ansible_version: "2.9"
+  platforms:
+    - name: Archlinux
+      versions:
+        - rolling
+  galaxy_tags:
+    - docker
+    - docker-compose
+    - systemd
+    - automation
+    - archlinux
+  repository: https://s.veen.world/cymais
+  issue_tracker_url: https://s.veen.world/cymaisissues
+  documentation: https://s.veen.world/cymais
 dependencies:
-  - system-maintenance-lock
+  - system-maintenance-lock

roles/heal-docker/tasks/main.yml

@@ -9,13 +9,14 @@
   copy:
     src: heal-docker.py
     dest: "{{heal_docker}}heal-docker.py"
+  notify: restart heal-docker.cymais.service
   when: run_once_heal_docker is not defined
 
 - name: create heal-docker.cymais.service
   template: 
     src: heal-docker.service.j2 
     dest: /etc/systemd/system/heal-docker.cymais.service
-  notify: reload heal-docker.cymais.service
+  notify: restart heal-docker.cymais.service
   when: run_once_heal_docker is not defined
 
 - name: set service_name to the name of the current role