Cause: Customer reports that masters in an HA environment with a configured proxy can not reach etcd hosts by DNS
Consequence: Openshift service will emit error "etcd cluster is unavailable or misconfigured"
Fix: Openshift Ansible installer updated to collect IP of all etcd hosts and insert them into the NO_PROXY settings for the HA masters
Result: Masters can now connect to etcd instances in HA environments with configured proxies
I've been looking at this one today. Attempting to set up an environment to test in which enables the atomic-openshift-master-controllers service. No branch yet for the work.
Working on this issue in here: https://github.com/openshift/openshift-ansible/pull/4806
Testing initiated just now.
Moving to POST until we merge this into release-3.6 branch
https://github.com/openshift/openshift-ansible/pull/5081 backport to release-3.6 branch
Verified with openshift-ansible-184.108.40.206.7-2.git.0.340aa2c.el7.noarch.rpm
The internal ips of etcd hosts are added correctly by default to NO_PROXY in /etc/sysconfig/atomic-openshift-master-controllers, and builds and deployments work well.
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.