Description of problem:
With oc cluster up, if you are running a dnsmasq locally (e.g. if using VPN to call the VPN's DNS only for queries in its domain) your cat /etc/resolv.conf looks like this
# Generated by NetworkManager
copying this into the kube-dns container running on "different localhost" is pretty much pointless as the dns is not running in there but on the actual host where oc cluster up was run.
As that record is effectively unusable, oc cluster up can do best effort work of replacing it with a public IP, the openshift cluster is configured with. User can specify --public-hostname flag to influence this and give us a hint.
Fix is here: https://github.com/openshift/origin/pull/20688
Pl validate in 4.1
No objection to taking a patch for 4.1, but I don't think this is a blocker. I adjusted the priority/severity to low to try and indicate that — if I should have changed just one or the other, please feel free to fix!
there is no `oc cluster up` in 4.1 anymore, this was reported and fixed for 3.11
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.