Introduction

I am attempting to upgrade my installation of Openshift-Ansible from 3.6 to a higher version.

Currently, I'm running the following in an attempt to upgrade to v3.7:

ANSIBLE_HOST_KEY_CHECKING=False ansible-playbook -i hosts -k openshift-ansible/playbooks/byo/openshift-cluster/upgrades/v3_7/upgrade.yml

I get the following failing task when running:

TASK [Check for invalid namespaces and SDN errors] ***********************************************************************************************************************************************************************
fatal: [10.0.0.51]: FAILED! => {"changed": false, "msg": "Failed to GET hostsubnet.", "results": {"cmd": "/usr/bin/oc get hostsubnet -o json -n default", "results": [{}], "returncode": 1, "stderr": "Unable to connect to the server: dial tcp: lookup docker1.foo.bar on 10.0.0.1:53: no such host\n", "stdout": ""}, "state": "list"}

Details

I've checked out the release-3.9 branch of the openshift-ansible project.

For the sake of brevity, I'll post only the portion of my hosts file which I think are relevant, please let me know if I have left out important details:

[OSEv3:children]
masters
nodes
etcd

openshift_master_cluster_method=native
openshift_master_cluster_hostname=10.0.0.51
openshift_master_cluster_public_hostname=10.0.0.51
osm_cluster_network_cidr=10.168.0.0/13

[masters]
10.0.0.51

[etcd]
10.0.0.51

[nodes]
10.0.0.51 openshift_node_labels="{'region': 'infra','zone': 'default','node-role.kubernetes.io/compute': 'true'}" openshift_schedulable=true
10.0.0.52 openshift_node_labels="{'region': 'infra','zone': 'default','node-role.kubernetes.io/compute': 'true'}"
10.0.0.53 openshift_node_labels="{'region': 'infra','zone': 'default','node-role.kubernetes.io/compute': 'true'}"

I think this error may be due to some sort of name resolution error, I have an /etc/resolv.conf inside 10.0.0.53 with the following contents:

nameserver 10.0.0.1
nameserver 10.0.0.53

If I do nslookup docker1.foo.bar 10.0.0.1, I get a failure to find, which is expected because 10.0.0.1 can't resolve internal network names.

If I do docker1.foo.bar 10.0.0.53, I resolve the name to 10.0.0.51 as expected.

What I've Already Done Before this

Initially, when I ran upgrade.retry as above, I got the following error:

TASK [openshift_excluder : Check the available origin-docker-excluder version is at most of the upgrade target version] **********************************************************************
[DEPRECATION WARNING]: Using tests as filters is deprecated. Instead of using `result|version_compare` use `result is version_compare`. This feature will be removed in version 2.9. 
Deprecation warnings can be disabled by setting deprecation_warnings=False in ansible.cfg.
fatal: [docker1.foo.bar]: FAILED! => {"changed": false, "msg": "Available origin-docker-excluder version 3.9.0 is higher than the upgrade target version"}
    ...

So, I've added the following to my inventory file:

enable_excluders=false

Which seems to make the playbook continue past this error, then I hit the error as described above in the Introduction.

Question

How can I upgrade my installation of Openshift-Ansible, or, what is causing my error?