Properly fail playbook when vault_restore.sh fails

Currently, when the main restore script vault_restore.sh fails, it goes
to the rescue block to delete the vault workdir. As ansible treats the
script block not as "failed", the playbook is not considered failed as
well. Now the rescue block is removed and vault workdir deletion is
moved under the always block, the playbook should properly fail when the
script as failed.

Test Plan:
PASS    The playbook fails with failed=1 when the script exits with 1
PASS    Vault workdir is deleted for both playbook failure and success

Story: 2011073
Task: 50971

Change-Id: Id6dd31852cc17da8f09d3e33bc7aae09d7969782
Signed-off-by: Tae Park <tae.park@windriver.com>

playbookconfig/src/playbooks/roles/vault/vault_restore/tasks/main.yml

@@ -59,16 +59,8 @@
       shell: >-
         kubectl exec -n "vault" "{{ vault_manager_pod_name.stdout }}" --
         bash -c "source /opt/script/init.sh; if [ -f $PAUSEFILE ]; then rm -f $PAUSEFILE; fi" 2>&1
-  rescue:
-    - name: Clean up vault subdir if in standalone mode
+    - name: Clean up vault subdir
       file:
         path: "{{ vault_backup_dir }}"
         state: absent
-      when: op_mode == "standalone"
       become: yes
-
-- name: Cleanup vault subdir
-  file:
-    path: "{{ vault_backup_dir }}"
-    state: absent
-  become: yes