A containerized ceph deployment which didn't pass any overrides but used only the defaults found in:
Produced a resources.WorkflowTasks_Step2_Execution Heat resource whose Mistral ansible-playbook call did not pass the osd_scenario parameter to the ansible-playbook --extra-vars run . This resulted in a overcloud deployment with 0 OSDs (0 OSD containers) though the deployment succeeded and evidence of the failure was not seen later until a tempest run.
Though ceph-ansible should better handle this scenario and fail as per BZ 1502841 , the defaults alone should work, provided that the ceph overcloud node has the device /dev/vdb.
Thanks John, as a side note, we might also default storage-environment.yaml to ceph-ansible at this point.
The root cause here is that /home/stack/virt/internal.yaml is overriding the default CephAnsibleDisksConfig passing only 'devices:', see 
Instead of closing this BZ though, we might use it to track migration of storage-environment.yaml from puppet-ceph to ceph-ansible if QE agrees with it being a good idea.
Yogev, Sasha, Marius, what do you guys think?
verified on openstack-tripleo-heat-templates-7.0.3-3.el7ost.noarch
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.