Start of master fails with error:
6109 start_master.go:103] could not init cloud provider "openstack": Post http://10.175.1.110:5000/v2.0/tokens: dial tcp 10.175.1.110:5000: i/o timeout
The IP address in the request is not correct,it is the keystone "internal" endpoint, it should be "public".
Why is "internal" endpoint coming to use with start_master.go, and how to enforce usage of "public" one?
When using HEAT template to create OpenShift, public endpoint was specified for --parameter os_auth_url
That public endpoint is not in any config file on the system (not in master config file, not in cloud-init conf files). It looks like the URL is somehow dynamically created, but from what?