When trying to use the new native OVSDB provider, the connectivity never goes up due to the fact that what seems to be the db_set operation failing to change the patch ports from "nonexistant-peer" to the correct peer, therefore not linking the bridges together. When using the following openvswitch agent configuration: [OVS] ovsdb_interface = ovsdb Restarting the OVS agent will setup everything but leave network in a failed state because the correct patch ports aren't updated: # ovs-vsctl show Bridge br-ex Port br-ex Interface br-ex type: internal Port "em1" Interface "em1" Port phy-br-ex Interface phy-br-ex type: patch options: {peer=nonexistent-peer} Bridge br-int fail_mode: secure Port "qvo25d28228-9c" tag: 1 Interface "qvo25d28228-9c" ... Port int-br-ex Interface int-br-ex type: patch options: {peer=nonexistent-peer} Configuration using ovs vsctl interface works with no problems. This issue renders ovsdb native interface unusable. It has been fixed in upstream and backported to stable/liberty branch. As there is not known plan for next stable RHEL OSP 8 release, the relevant patches are going to be backported proactively.
Why do we care about the issue in OSP context? Do we use native interface anywhere?
Yes, in functional tests. Namely functional tests for bug 1340717 fail if this patch is not included.
Code tested in latest OSP8 - openstack-neutron-7.0.4-7.el7ost
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-2016:1353