switch_to_containers: use ceph binary from container

use the ceph binary from the container instead of the host.
If the ceph CLI version isn't compatible between host and container
image, it can cause the CLI to hang.

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

diff --git a/infrastructure-playbooks/switch-from-non-containerized-to-containerized-ceph-daemons.yml b/infrastructure-playbooks/switch-from-non-containerized-to-containerized-ceph-daemons.yml
index b0c49d84cfc8d07e9af994465bb6fe91eccc55ee..987cc1b42d9d9bf50ae279eaa713487cee95cead 100644 (file)
--- a/infrastructure-playbooks/switch-from-non-containerized-to-containerized-ceph-daemons.yml
+++ b/infrastructure-playbooks/switch-from-non-containerized-to-containerized-ceph-daemons.yml
@@ -129,16 +129,13 @@
         name: ceph-mon
 
   post_tasks:
-    # We don't do a container test by running 'docker exec ...' since not all the monitors have switched to containers yet.
-    # Thus, we continue to use the 'ceph' binary from the host, there is no issue with that.
-    - name: non container | waiting for the monitor to join the quorum...
-      command: ceph --cluster "{{ cluster }}" -s --format json
+    - name: waiting for the monitor to join the quorum...
+      command: "{{ docker_exec_cmd }} ceph --cluster {{ cluster }} -s --format json"
       register: ceph_health_raw
       until: >
         hostvars[mon_host]['ansible_hostname'] in (ceph_health_raw.stdout | from_json)["quorum_names"]
       retries: "{{ health_mon_check_retries }}"
       delay: "{{ health_mon_check_delay }}"
-      delegate_to: "{{ mon_host }}"
 
 - name: switching from non-containerized to containerized ceph mgr