Description of problem:
Installation failed on AWS while enabling cloudprovider. Dig more, found that the root cause of atomic-openshift-node failure was `AWS_ACCESS_KEY_ID` and `AWS_SECRET_ACCESS_KEY` not set in /etc/sysconfig/atomic-openshift-node.

TASK [openshift_node : Abort if node failed to start] **************************
Friday 29 September 2017  09:37:13 +0000 (0:00:01.155)       0:19:38.142 ****** 
fatal: [ec2-54-89-99-146.compute-1.amazonaws.com]: FAILED! => {"changed": false, "failed": true, "msg": "Node failed to start please inspect the logs and try again"}

Check the logs:
Sep 29 05:36:46 ip-172-18-11-50.ec2.internal atomic-openshift-node[30618]: I0929 05:36:46.462338   30627 aws.go:806] Building AWS cloudprovider
Sep 29 05:36:46 ip-172-18-11-50.ec2.internal atomic-openshift-node[30618]: F0929 05:36:46.464669   30627 start_node.go:141] could not init cloud provider "aws": error finding instance i-054489e66654e2cc8: "error listing AWS instances: \"NoCredentialProviders: no valid providers in chain. Deprecated. \\n\\tFor verbose messaging see aws.Config.CredentialsChainVerboseErrors\""
Sep 29 05:36:46 ip-172-18-11-50.ec2.internal systemd[1]: atomic-openshift-node.service: main process exited, code=exited, status=255/n/a
"log" 6048L, 975083C

[root@ip-172-18-11-50 ~]# cat /etc/sysconfig/atomic-openshift-node
OPTIONS=--loglevel=2
CONFIG_FILE=/etc/origin/node/node-config.yaml
IMAGE_VERSION=v3.7.0

Version-Release number of the following components:
ansible 2.3
openshift-ansible-3.7.0-0.128.0.git.0.89dcad2.el7.noarch.rpm

How reproducible:
sometimes

Steps to Reproduce:
1. Trigger installation against AWS with cloudprovicer enabled.
2.
3.

Actual results:
See above

Expected results:

Additional info:
This issue was randomly happen due to https://github.com/ansible/ansible/issues/24450

Introduced by https://github.com/openshift/openshift-ansible/pull/5230 that the task "Start and enable node" is executed prior to `Configure AWS Cloud Provider Settings`.