We have this exact setup.  You are correct, you will need Virtual IP's for each 
public WAN IP that OVH have assigned you.  We have separate services listening 
on x.x.x.1, x.x.x.2, x.x.x.3 etc, works like a charm.

JC

-----Original Message-----
From: List [mailto:[email protected]] On Behalf Of Adam Thompson
Sent: August-01-17 12:57 PM
To: [email protected]
Subject: [pfSense] IPv6 problem at OVH

Wondering how anyone else manages (or would manage) this scenario:

* Private Cloud at OVH.  (Runs VMware, which isn't terribly relevant
AFAICT.)
* OVH provides a single VLAN that is connected directly to their router
* ALL public IP addresses are terminated on that VLAN (i.e. bound directly to 
that interface on their router) including the entire IPv6 /56.
*** As a consequence, all IPv4 addresses must respond to ARP, and all
IPv6 addresses must respond to NDP, in order to be successfully publicly routed.
(And yes, they gave me an entire /56 of IPv6... that isn't routed or broken up 
in any way.  And they won't subnet or route anything to me.  
Yay.)
* Meanwhile, I have public services (multiple tenants) running on multiple 
VLANs, each behind a single pfSense firewall with a WAN interface in the 
massive public-address-space VLAN.
* I very much want the service address to be different from the firewall 
address, i.e. the firewall WAN i/f might be bound to 1.2.3.4, then I want the 
publicly-accessible service to live at 1.2.3.5, so that I can distinguish based 
on reverse DNS whether outbound connections are coming from the firewall or 
from the customer's server.  This works great with IPv4, a Proxy ARP VIP, and 
1:1 NAT.
* I also need to provide IPv6 connectivity inbound AND outbound, ideally with 
the same reverse-dns differentiation.

I've tried 1:1 NAT, which seems to break IPv6 altogether every time I configure 
it (although JimP can't reproduce it yet, so presumably it's somehow 
environment-specific).  I'm unclear whether this will work anyway with the NDP 
adjacency requirement.

I've tried NPt, which doesn't do NDP, and so doesn't work in this scenario.

The next thing I can try (but haven't yet) is an IP Alias VIP with Port 
Forwarding, and then... maybe a custom Outbound NAT rule?

Am I missing something fundamental?  I know what OVH is doing is stupid (NDP 
for an entire /56?  Fee fi fo fum, I smell a DoS attack...) , but they have 
2000+ other customers on this exact platform, surely ONE of them must have a 
similar situation!  I know IPv6 is new, but ... surely one them must run IPv6?

Again: IPv4 isn't a problem because Proxy ARP works great and solves the 
silliness of them not routing those allocated subnets to me.  IPv6 is a problem 
because pfSense has to handle NDP *and* do NAT and I can't find a way to make 
it do that properly


Thoughts/opinions/brickbats welcome.
-Adam

P.S. I seem to not be receiving emails from the list reliably, kindly CC me if 
you don't mind...
_______________________________________________
pfSense mailing list
https://lists.pfsense.org/mailman/listinfo/list
Support the project with Gold! https://pfsense.org/gold
_______________________________________________
pfSense mailing list
https://lists.pfsense.org/mailman/listinfo/list
Support the project with Gold! https://pfsense.org/gold

Reply via email to