[federation] Fix race condition with openstackclient pod readiness

After Keystone reconfiguration the openstackclient pod may restart,
causing oc exec commands to fail with "pods openstackclient not found".
Add a readiness wait before running OpenStack setup commands, and add
retries to run_osp_cmd.yml matching the pattern used by the ipa role.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Signed-off-by: Luca Miccini <lmiccini@redhat.com>

Author: 2 people

roles/federation/tasks/hook_post_deploy.yml

@@ -30,6 +30,22 @@
   retries: 30
   delay: 20
 
+- name: Wait for openstackclient pod to be ready
+  kubernetes.core.k8s_info:
+    kubeconfig: "{{ cifmw_openshift_kubeconfig }}"
+    api_version: v1
+    kind: Pod
+    name: openstackclient
+    namespace: "{{ cifmw_federation_run_osp_cmd_namespace }}"
+  register: _openstackclient_pod
+  until: >-
+    _openstackclient_pod.resources | length > 0 and
+    (_openstackclient_pod.resources[0].status.phase | default('') == 'Running') and
+    (_openstackclient_pod.resources[0].status.containerStatuses | default([]) |
+    selectattr('ready', 'equalto', true) | list | length > 0)
+  retries: 30
+  delay: 10
+
 - name: Build realm configurations for single realm OpenStack setup
   ansible.builtin.set_fact:
     _federation_openstack_realms_to_process:

roles/federation/tasks/run_osp_cmd.yml

@@ -26,3 +26,5 @@
       --
       {{ _osp_cmd }}
   register: federation_run_ocp_cmd
+  retries: 10
+  delay: 10