Routed networks and FIP gateway assignments

Syd Logan syd.logan at
Wed Apr 17 00:51:33 UTC 2019

I'm currently learning about routed networks and have a lab that is
implementing a 3 tier leaf spine and have a question.

I know that I can create a subnet from which FIP gateway IP addresses are
consumed, e.g.,

openstack subnet create demo-floating-ip-agent-gateway-subnet \
  --subnet-range --no-dhcp \
  --service-type 'network:floatingip_agent_gateway' \
  --network demo-ext-net

In videos from past OpenStack summits such as this drawings show that the subnet is used within a
single segment's compute nodes, and I can believe that would work. What
isn't clear is whether the same subnet should be used across multiple
segments, or the intent is for each segment defined in the routed network
to have a unique subnet from which these FIP gateway IPs are assigned,
e.g., some other segment would be using, say

In the non-routed network setup where I leanred about DVR it was pretty
clear only one subnet was allowed, but what is the best practice/intent for
routed nets?


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the openstack-discuss mailing list