I've spent the last two days trying to reproduce these problems on an OpenStack Cloud with three gateway/network units, each with 8 cores and 16G of RAM; I have 99 routers configured on the cloud, plumbed into a shared external provider network, and I've been switching them between ha and non-ha configuration continually for the last 24 hours using:
openstack router set --disable $router_id openstack router set --{no-}ha $router_id openstack router set --enable $router_id Routers successfully switched between HA and non-HA mode, with keepalived, netns and ovs ports being tidied as appropriate. We think we saw issues with ovs performance when the db contained a large number of dangling ovs interfaces on br-int, so I loaded on of my units up with 400 dangling ovs instances. I'll run the switching again for a while and see if that exacerbates things sufficiently to cause the same issue again. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1749425 Title: Neutron integrated with OpenVSwitch drops packets and fails to plug/unplug interfaces from OVS on router interfaces at scale To manage notifications about this bug go to: https://bugs.launchpad.net/neutron/+bug/1749425/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs