+++ This bug was initially created as a clone of Bug #1513764 +++ Currently egress-routers always attach themselves to the network interface containing the default route. We need to make it possible to attach one to an alternate interface. To get this out to a customer quickly, we're doing the simplest possible fix, which is to let the user manually specify a network interface (rather than, say, having OpenShift try to figure out the correct interface automatically based on EGRESS_SOURCE).
https://github.com/openshift/ose/pull/929
Merged... how do we get a hotfix built?
(In reply to Dan Winship from comment #1) > https://github.com/openshift/ose/pull/929 (plus https://github.com/openshift/ose/pull/936 to fix the build afterward)
verified in atomic-openshift-3.6.173.0.83-1.git.0.1f70b16.el7.x86_64 and passed.
Just for reference: The annotation will still accept true and false pod.network.openshift.io/assign-macvlan: "true" this allows it to accept further values like pod.network.openshift.io/assign-macvlan: "eth1"
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. https://access.redhat.com/errata/RHBA-2017:3438