Refactor systemd handling

- sys-ctl-rpr-btrfs-balancer: suppress service flush for btrfs balancer (too expensive to run each play)
- sys-daemon: replace raw systemctl calls with ansible.builtin.systemd (daemon_reload, daemon_reexec)
- sys-service: split handler into 'Enable systemctl service' and 'Set systemctl service state', add become, async/poll, suppress flush guard

Conversation: https://chatgpt.com/share/68c2f7a6-6fe4-800f-9d79-3e3b0ab4a563

roles/sys-ctl-rpr-btrfs-balancer/tasks/01_core.yml

@@ -12,6 +12,7 @@
 - include_role:
     name: sys-service
   vars:
+    system_service_suppress_flush:  true # It takes a super long time - Better wait for failure of timed service instead of executing it on every play
     system_service_copy_files:      false
     system_service_on_calendar:     "{{ SYS_SCHEDULE_REPAIR_BTRFS_AUTO_BALANCER }}"
     system_service_timer_enabled:   true