Ask is moving to Stack Overflow and! Please use the "opendaylight" tag on either of these sites. This site is now in Read-Only mode

Revision history [back]

click to hide/show revision 1
initial version

So it seems like the solution to this problem is to manually set up bridge connections between the OVS switches. I loosely followed the instructions here:

In my case the code looks like:

read ovstbl <<< $(ovs-vsctl get Open_vSwitch . _uuid)
bridgeip=$(ip -o addr show | grep 10.11.10 | grep -w inet | awk '{print $4}' | sed -e 's/\/.*//g')
ovs-vsctl set Open_vSwitch $ovstbl other_config={'local_ip'=$bridgeip}

Then on my network manager node, the following commands connect us to the external public internet:

externaliface=`ip route show default | grep default | cut -d' ' -f5`
ovs-vsctl add-br br-ex
ovs-vsctl add-port br-ex $externaliface

However this code is very specific to my environment. Our network manager node is isolated from the controller and compute nodes, I doubt this would work on a single-node deployment. We have a management VLAN running on 10.11.10.* addresses, so I've hard-coded this subnet to determine the bridge IP. Moreover any OpenStack/Neutron networks we use must be of GRE type, or else instances will still have network problems.

So at best this is a hacky, partial solution. If anybody has something that works more generally, please post it here.