Replace block/rescue with ignore_errors/register

Users are always getting triggered by the failed
`systemctl reload-or-restart` command. To potentially reduce confusion,
let's replace block/rescue back to `ignore_errors`, as in this case
Ansible at least printing out `...ignoring` for the raised error,
instead of silently proceeding further, which might help out
to see that the failure is non-critical.

Change-Id: I264e23598dd211c5ddef971d2cc3702f4f2af3b6

tasks/systemd_mounts.yml

@@ -83,18 +83,20 @@
 #                  the task will be rescued and the regular
 #                  systemd module will be attempted before
 #                  failing the task run.
-- name: Mount state block
-  block:
-    - name: Set the state of the mount  # noqa: command-instead-of-module
-      ansible.builtin.command: "systemctl {{ systemd_mount_states[item.state] }} {{ mount_service_name.stdout }}"
-      changed_when: false
-      when:
-        - item.state is defined
-  rescue:
-    - name: Set the state of the mount (fallback)
-      ansible.builtin.systemd:
-        name: "{{ mount_service_name.stdout }}"
-        state: "{{ item.state }}"
+- name: Set the state of the mount  # noqa: command-instead-of-module
+  ansible.builtin.command: "systemctl {{ systemd_mount_states[item.state] }} {{ mount_service_name.stdout }}"
+  changed_when: false
+  ignore_errors: true
+  register: set_mount_state_command
+  when:
+    - item.state is defined
+
+- name: Set the state of the mount (fallback)
+  ansible.builtin.systemd:
+    name: "{{ mount_service_name.stdout }}"
+    state: "{{ item.state }}"
+  when:
+    - set_mount_state_command is failed
 
 - name: Unload mount(s)
   ansible.builtin.systemd: