A very clever user on #openstack-ansible irc pointed out that the command was executed as a local_action. hence, it was executed on the deployment host - and not on the glance containers as I expected.
It turned out that i has a routing problem with my ansible deployment host, which could not access the ceph mon hosts.
This issue may be closed. However, a helpfull error text describing that the deployment host is not able to access the ceph mons would be nice.
A very clever user on #openstack-ansible irc pointed out that the command was executed as a local_action. hence, it was executed on the deployment host - and not on the glance containers as I expected.
It turned out that i has a routing problem with my ansible deployment host, which could not access the ceph mon hosts.
This issue may be closed. However, a helpfull error text describing that the deployment host is not able to access the ceph mons would be nice.
- name: Verify Ceph monitors are up ile=/dev/ null ecking= no
local_action: >
command ssh
-o ConnectTimeout=6
-o UserKnownHostsF
-o StrictHostKeyCh
{{ item }}
exit