Description of problem:

When "openshift-master/scaleup.yml" runs for additional masters, always it labeled "compute" role to new master added badly.

The testing evidence for adding "master-3.ocp.example.com" as follows.

* Inventory file
~~~
[OSEv3:children]
masters
nodes
etcd
new_nodes
new_masters

...
[new_nodes]
master-3.ocp.example.com

[new_masters]
mastre-3.ocp.example.com
~~~

* Before running "openshift-master/scaleup.yml"
~~~
# oc get node
NAME                       STATUS    ROLES            AGE       VERSION
infra-1.ocp.example.com    Ready     infra            5h        v1.9.1+a0ce1bc657
master-1.ocp.example.com   Ready     master           5h        v1.9.1+a0ce1bc657
master-2.ocp.example.com   Ready     master           5h        v1.9.1+a0ce1bc657
node-1.ocp.example.com     Ready     compute          5h        v1.9.1+a0ce1bc657
~~~

* After running "openshift-master/scaleup.yml"
~~~
# oc get node
NAME                       STATUS    ROLES            AGE       VERSION
infra-1.ocp.example.com    Ready     infra            5h        v1.9.1+a0ce1bc657
master-1.ocp.example.com   Ready     master           5h        v1.9.1+a0ce1bc657
master-2.ocp.example.com   Ready     master           5h        v1.9.1+a0ce1bc657
master-3.ocp.example.com   Ready     compute,master   4m        v1.9.1+a0ce1bc657
node-1.ocp.example.com     Ready     compute          5h        v1.9.1+a0ce1bc657
~~~


Version-Release number of the following components:
rpm -q atomic-openshift-utils
atomic-openshift-utils-3.9.74-1.git.0.70a0a63.el7.noarch

rpm -q ansible
ansible-2.6.11-1.el7ae.noarch

ansible --version
ansible 2.6.11
  config file = /etc/ansible/ansible.cfg
  configured module search path = [u'/root/.ansible/plugins/modules', u'/usr/share/ansible/plugins/modules']
  ansible python module location = /usr/lib/python2.7/site-packages/ansible
  executable location = /usr/bin/ansible
  python version = 2.7.5 (default, Sep 12 2018, 05:31:16) [GCC 4.8.5 20150623 (Red Hat 4.8.5-36)]

How reproducible:

If you add one new master node using "openshift-master/scaleup.yml", it always happens.

Steps to Reproduce:
1.
2.
3.

Actual results:
"compute,master" node-role label is added to new master node.

Expected results:
"master" node-role label is added to new master node.

Additional info:

The following task "Label all-in-one master as a compute node" labeled "compute" badly to new master.

~~~
TASK [openshift_manage_node : Label all-in-one master as a compute node] *********************************************************************************************************************
task path: /usr/share/ansible/openshift-ansible/roles/openshift_manage_node/tasks/set_default_node_role.yml:23
Using module file /usr/share/ansible/openshift-ansible/roles/lib_openshift/library/oc_label.py
<10.0.1.10> ESTABLISH SSH CONNECTION FOR USER: quicklab
<10.0.1.10> SSH: EXEC ssh -C -o ControlMaster=auto -o ControlPersist=900s -o GSSAPIAuthentication=no -o StrictHostKeyChecking=no -o KbdInteractiveAuthentication=no -o PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey -o PasswordAuthentication=no -o User=quicklab -o ConnectTimeout=10 -o ControlPath=/home/quicklab/.ansible/cp/9ac3eceffa 10.0.1.10 '/bin/sh -c '"'"'sudo -H -S -n -u root /bin/sh -c '"'"'"'"'"'"'"'"'echo BECOME-SUCCESS-kuczpnvvhprzshkcozwtwbuvijoktkbr; /usr/bin/python'"'"'"'"'"'"'"'"' && sleep 0'"'"''
<10.0.1.10> (0, '\n{"invocation": {"module_args": {"kind": "node", "name": "master-3.ocp.example.com", "labels": [{"value": "true", "key": "node-role.kubernetes.io/compute"}], "namespace": null, "kubeconfig": "/etc/origin/master/admin.kubeconfig", "state": "add", "debug": false, "selector": null}}, "state": "add", "changed": true, "results": {"returncode": 0, "cmd": "/bin/oc label node master-3.ocp.example.com node-role.kubernetes.io/compute=true --overwrite", "results": {}}}\n', '')
changed: [10.0.1.9 -> 10.0.1.10] => {
    "changed": true, 
    "failed": false, 
    "invocation": {
        "module_args": {
            "debug": false, 
            "kind": "node", 
            "kubeconfig": "/etc/origin/master/admin.kubeconfig", 
            "labels": [
                {
                    "key": "node-role.kubernetes.io/compute", 
                    "value": "true"
                }
            ], 
            "name": "master-3.ocp.example.com", 
            "namespace": null, 
            "selector": null, 
            "state": "add"
        }
    }, 
    "results": {
        "cmd": "/bin/oc label node master-3.ocp.example.com node-role.kubernetes.io/compute=true --overwrite", 
        "results": {}, 
        "returncode": 0
    }, 
    "state": "add"
}
~~~