Description of problem:
Trigger an scaleup and input new node hostname. When it comes to ensure facts, input 'N' to exit from the interaction due to the gathered hostname for the new node is not correct, then edit installer.cfg.yml to update the new added node info. But after that, when run scaleup again with '-c', it can not detect this fresh node and still ask user to input new node info again. 

Version-Release number of selected component (if applicable):
atomic-openshift-utils-3.4.13-1.git.0.ff1d588.el7.noarch


How reproducible:
always

Steps to Reproduce:
1.Trigger an scaleup
# atomic-openshift-installer scaleup

2.Edit installer.cfg.yml to add an label for the new node.
#cat /root/.config/openshift/installer.cfg.yml
<--snip-->
deployment:
  ansible_ssh_user: root
  hosts:
 ....
  - connect_to: openshift-139.x.x.x
    hostname: 192.168.2.186
    ip: 192.168.2.186
    node_labels: '{''region'': ''infra''}'
    public_hostname: 10.66.147.139
    public_ip: 10.66.147.139
    roles:
    - node
<--snip-->

3.Run scaleup again
# atomic-openshift-installer -c /root/.config/openshift/installer.cfg.yml scaleup

Actual results:
It will print installation summary and then go to the workflow "New Node Configuration" and ask user to input new node's info again.It seems that it does not detect which hosts are installed and which hosts are fresh. It will result that user can not edit info in installer.cfg.yml when scaleup.

Expected results:
It will detected fresh hosts and pass "New Node Configuration" step to go to left steps for scaleup.