rgw: add a retry on pool related tasks

sometimes those tasks might fail because of a timeout.
I've been facing this several times in the CI, adding this retry might
help and won't hurt in any case.

Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>

roles/ceph-rgw/tasks/main.yml

@@ -31,6 +31,8 @@
       changed_when: false
       with_dict: "{{ rgw_create_pools }}"
       delegate_to: "{{ groups[mon_group_name][0] }}"
+      register: result
+      until: result is succeeded
       run_once: true
 
     - name: customize pool size
@@ -39,5 +41,7 @@
       delegate_to: "{{ groups[mon_group_name][0] }}"
       changed_when: false
       run_once: true
+      register: result
+      until: result is succeeded
       when:
         - item.size | default(osd_pool_default_size) != ceph_osd_pool_default_size