OVN L3 is currently a single point of failure and it should be highly available.
Discussing up with the community https://mail.openvswitch.org/pipermail/ovs-dev/2017-April/330597.html And talking about limitations of the planned approach to do L3 HA Active/Standby, which are also present on VRRP or other protocols. https://mail.openvswitch.org/pipermail/ovs-dev/2017-April/330597.html
This will require OVS 2.8, pushed to OSP 13.
ovn-core patches merged, now we need to work on the networking-ovn side.
An idea we just discussed was extending the northbound database with a simple boolean on the Gateway_Chassis table to reflect if it's the current master or not. That information is already available through the southbound database by examining the Port_Binding table, but doing something in the northbound db would be a usability improvement. One catch is that the current db schema allows you to reference a single Gateway_Chassis for multiple gateways, so maybe putting the column on Gateway_Chassis doesn't work. Another option could be a new column on Logical_Router_Port that references a single gateway_chassis row to indicate which is the current master.
*** Bug 1498109 has been marked as a duplicate of this bug. ***
verified on OSP13 puddle 2018-04-03.3 openvswitch-2.9.0-15.el7fdp.x86_64 there are some bugs. All the results updated in test run: https://polarion.engineering.redhat.com/polarion/#/project/RHELOpenStackPlatform/testrun?id=20180327-1217
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/RHEA-2018:2086