README.md

systemd service manager Ansible role

This is an Ansible role which manages systemd services.

Features

• starting (restarting) services, in order, according to their priority. Services can all be stopped cleanly and then started anew, restarted one-by-one, or started in priority batches without blocking (see devture_systemd_service_manager_service_restart_mode)

• making services auto-start (see devture_systemd_service_manager_services_autostart_enabled)

• verifying services managed to start (see devture_systemd_service_manager_up_verification_enabled)

• stopping services, in order, according to their priority

• starting/stopping all defined services, or a group of services (--tags=restart-group, --tags=stop-group)

• restarting services by cleanly stopping them and restarting them, one by one, or by priority batches

Usage

Example playbook:

- hosts: servers
  roles:
    - when: devture_systemd_service_manager_enabled | bool
      role: galaxy/com.devture.ansible.role.systemd_service_manager

Example playbook configuration (group_vars/servers or other):

# See `devture_systemd_service_manager_services_list_auto` and `devture_systemd_service_manager_services_list_additional`
devture_systemd_service_manager_services_list_auto: |
  {{
    ([{'name': 'some-service.service', 'priority': 1000}])
    +
    ([{'name': 'another-service.service', 'priority': 1500}])
  }}

Example playbook invocations tags (e.g. ansible-playbook -i inventory/hosts setup.yml --tags=XXXXX):

• restart, restart-all, start-all - restarts all services and potentially makes them auto-start (depending on devture_systemd_service_manager_services_autostart_enabled)

• restart-group, start-group - restarts services belonging to the specified group (e.g. --extra-vars="group=core")

• stop, stop-all - stops all services

• stop-group - stops services belonging to the specified group (e.g. --extra-vars="group=core")

Restart mode options for devture_systemd_service_manager_service_restart_mode:

• all-at-once (default) - issues a single systemctl restart (or start) command with all service names, letting systemd handle the entire transaction. See defaults/main.yml for detailed technical rationale
• clean-stop-start - stops all services in reverse priority order, then starts them in priority order
• one-by-one - restarts each service in priority order
• priority-batched - starts/restarts services in priority batches and queues them without blocking inside each batch

All-at-once retry controls:

• devture_systemd_service_manager_all_at_once_retries (default: 1) - passed directly to task retries for each single all-at-once systemctl restart ... or systemctl start ... command; must be an integer >= 1 (numeric strings accepted)
• devture_systemd_service_manager_all_at_once_retry_delay_seconds (default: 10) - delay between attempts; must be a non-negative integer (numeric strings accepted)

For a detailed comparison of these modes with real-world downtime benchmarks, see Restart Mode Comparison.