this tasks isn't using the right container_exec_cmd, that's delegating
to the wrong node.
Let's use the right fact to fix this command.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
(cherry picked from commit
ec33ee7574172c6f946f1030c26b8b4a6eb285dc)
when: dashboard_enabled | bool
- name: wait for all mgr to be up
- shell: "{{ container_exec_cmd | default('') }} ceph --cluster {{ cluster }} mgr dump -f json | python -c 'import sys, json; print(json.load(sys.stdin)[\"available\"])'"
+ shell: "{{ container_exec_cmd_mgr | default('') }} ceph --cluster {{ cluster }} mgr dump -f json | python -c 'import sys, json; print(json.load(sys.stdin)[\"available\"])'"
register: mgr_dump
retries: 30
delay: 5