Steps to reproduce
==================
Second deployement fails if using same working directory, Example:-
1) bash quickstart.sh -R master --tags all $VIRTHOST
once it's successful, 2nd run
2) bash quickstart.sh -R master --tags all $VIRTHOST
Above 2nd run of quickstart.sh fails at:-
TASK [tripleo-inventory : create inventory from template] **********************
task path: /home/centos/.quickstart/tripleo-quickstart/roles/tripleo-inventory/tasks/inventory.yml:126
Friday 15 December 2017 13:35:19 +0530 (0:00:00.165) 0:00:42.770 *******
changed: [localhost -> localhost] => {"changed": true, "checksum": "d90112ddb61d7f5a1f3d3ede9dc0d037407b0d0d", "dest": "/home/centos/.quickstart/hosts", "failed": false, "gid": 1002, "group": "centos", "md5sum": "2afb8ae2737ae5d2ef3d6ca25973a0f2", "mode": "0664", "owner": "centos", "secontext": "unconfined_u:object_r:user_home_t:s0", "size": 359, "src": "/home/centos/.ansible/tmp/ansible-tmp-1513325119.82-66400531597808/source", "state": "file", "uid": 1002}
TASK [tripleo-inventory : regenerate ssh config] *******************************
task path: /home/centos/.quickstart/tripleo-quickstart/roles/tripleo-inventory/tasks/inventory.yml:132
Friday 15 December 2017 13:35:20 +0530 (0:00:00.634) 0:00:43.405 *******
fatal: [localhost -> localhost]: FAILED! => {"changed": false, "failed": true, "msg": "AnsibleUndefinedVariable: \"hostvars['undercloud']\" is undefined"}
Workaround:-
Either remove default workdir ~/.quickstart or use another working directory by specifying -w <working dir>
This is likely caused by: https://review.openstack.org/#/c/525237/
Expected result
===============
Second run of quickstart should succeed.
Out of curiosity, what is the status of this? I was hit with https:/