`208.80.155.128/25` is currently statically routed to `10.64.20.13` (`labnet1001`) and used for [[ https://docs.openstack.org/ocata/user-guide/cli-manage-ip-addresses.html | floating-ip ]]'s in the cloud setup.
DNS for these is delegated to the nameservers managed by [[ https://docs.openstack.org/designate/latest/ | designate ]] as well:
> templates/155.80.208.in-addr.arpa
```; 208.80.155.128/25 Eqiad Labs virtualization subnet
; Delegate 208.80.155.128 - 208.80.155.255 to labs-ns*
128-25 IN NS labs-ns0.wikimedia.org.
IN NS labs-ns1.wikimedia.org.```
Out of this `/25` (128 possible hosts) we are currently using 104 addresses
nova list --all-tenants | grep 208
{P7059, lines=10}
This is misleading actually as our policy in Toolforge is for execs to have their own IPs and that isn't being honored so we could/should be using up another 20 or so I expect.
----
Request: `/24` range of publicv4 IPs that can used for SNAT/DNAT in the Neutron deployment in eqiad.
Compensating controls: The /25 in use now can be returned to the general pool as soon as we move to the new range (expectation is late calendar year 2018). Our plans are predicated upon being able to have both the existing nova-network and the new Neutron based deployments existing concurrently for comprehensive testing and smooth transition. We could allocate another like /25 and get by for now but my thinking here is a /24 gets 2-3 years down the road at least. Total judgement call there based on current availability.