While using -e ceph_ansible_limit to exclude nodes the task gather uuid task [0] still runs on all nodes. 

This can affect deployments at scale when a couple of nodes are offline because the operator won't be able to exclude them using ansible's --limit feature.

The problem is that the ceph_ansible_limit is getting passed to the ceph-ansible run but not to the nodes-uuid run [2]. 

Pasting line 55 of [1] into line 32 of [2] should give you a quick workaround.

[0] https://github.com/openstack/tripleo-ansible/blob/ca165ec94018293dfbbbd37c44a2704b8144f203/tripleo_ansible/roles/tripleo_ceph_uuid/tasks/gather.yml#L17 

[1] https://github.com/openstack/tripleo-ansible/blob/stable/train/tripleo_ansible/roles/tripleo-ceph-run-ansible/tasks/main.yml#L55
[2] https://github.com/openstack/tripleo-ansible/blob/ca165ec94018293dfbbbd37c44a2704b8144f203/tripleo_ansible/roles/tripleo_ceph_uuid/tasks/gather.yml#L17