Red Hat Bugzilla – Bug 1299065
Fix etcd hostname/ip address handling
Last modified: 2016-07-03 20:47:18 EDT
Description of problem:
When the openshift_ip and openshift_hostname variables are defined in the /etc/ansible/hosts file, they appear to only be used for the installation and then not enforced anywhere else.
Version-Release number of selected component (if applicable):
Openshift 3.1 installer
Steps to Reproduce:
1. Setup /etc/ansible/hosts file to use a private interface, excluding the public interface
2. Run the installer script
The installer uses the private interface but then the cluster fails to get started because it tries, and fails, to use the public interface.
The installer and Cluster use the private interface as setup in the ansible hosts file
We should wrap the etcd role(s) with openshift_etcd role(s) so that we can provide a consistent interface for configuring things like hostnames and ips for the etcd hosts.
Tracking this in the following trello card: https://trello.com/c/2tygIgBm
Is this something that will be corrected with the certificate and delegate_to refactoring you've been working on?
(In reply to Jason DeTiberus from comment #11)
> Is this something that will be corrected with the certificate and
> delegate_to refactoring you've been working on?
No, the delegate_to isn't taking this into account currently.
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.