Comment 2 for bug 1599329

The cause of this is that multiple ports are getting the same ofport assignment in OVS. This can be seen the vswitchd log:

2016-07-05T22:35:33.365Z|01321|netdev_linux|INFO|ioctl(SIOCGIFHWADDR) on tap72cf6622-f4 device failed: No such device
2016-07-05T22:35:34.508Z|01322|bridge|INFO|bridge br-int: added interface tapc576ea66-4a on port 372
2016-07-05T22:35:38.347Z|01323|bridge|INFO|bridge br-int: added interface qvocec23074-06 on port 372
2016-07-05T22:35:38.355Z|01324|netdev_linux|INFO|ioctl(SIOCGIFHWADDR) on tapc576ea66-4a device failed: No such device
2016-07-05T22:35:41.467Z|01325|bridge|INFO|bridge br-int: added interface tapc576ea66-4a on port 372
2016-07-05T22:35:41.483Z|01326|bridge|INFO|bridge br-int: added interface tapf227de76-76 on port 378
2016-07-05T22:35:41.683Z|01327|netdev_linux|INFO|ioctl(SIOCGIFHWADDR) on tapf227de76-76 device failed: No such device
2016-07-05T22:35:42.608Z|01328|bridge|INFO|bridge br-int: added interface qvocec23074-06 on port 372
2016-07-05T22:35:42.621Z|01329|netdev_linux|INFO|ioctl(SIOCGIFHWADDR) on tapc576ea66-4a device failed: No such device
2016-07-05T22:35:43.171Z|01330|bridge|INFO|bridge br-int: added interface tapc576ea66-4a on port 372
2016-07-05T22:35:46.147Z|01331|bridge|INFO|bridge br-int: added interface qvocec23074-06 on port 372

My suspicion is that this is a direct result of us ignoring the guideline not to use OVS devices that get moved into namespaces that was outlined in http://lists.openstack.org/pipermail/openstack-dev/2015-February/056834.html and https://bugzilla.redhat.com/show_bug.cgi?id=1160340