I am colocating some hardware in a local datacenter, and I'm trying to get IPv6 going. My datacenter has assigned me a /64, but they have given me a gateway address within that /64.
The gateway (that I don't control) is 2606:1000:1000:1::1
I have two uplinks into the datacenter, which go to (my) two Ubiquiti Edgerouters that I control. Eventually, I'd like to get VRRP going, but for now, I just have some of my IP space on 1 router, and some on the other.
So on one of the edgerouters, I've configured the following:
- eth0 = 2606:1000:1000:1::11/64
- eth2 = 2606:1000:1000:1::111/64
- static route is set like so:
set protocols static route6 '::/0' next-hop '2606:1000:1000:1::1' interface eth0
- I've assigned 1 of my hosts the following IP: 2606:1000:1000:1::210/64
My symptoms are as follows:
- The router can ping6 google.com and get responses back.
- The host can ping6 2606:1000:1000:1::111 and get responses back.
- The host cannot ping6 google.com.
- The router can NOT ping6 the host at 2606:1000:1000:1::210
I feel fairly positive that the router's set protocols static route6 is causing the issue here, but I'm having trouble figure out how to get a static route setup for the /64 to hairpin it back into eth2.
The following attempt doesn't work, when I try to configure a static route for my own /64:
cha-rtr1# set protocols static route6 '2606:1000:1000:1::0/64' next-hop '2606:1000:1000:1::111' interface eth2
[edit]
cha-rtr1# commit
[ protocols static route6 2606:1000:1000:1::0/64 next-hop 2606:1000:1000:1::111 ]
Error: Nexthop address cannot be same as own interface address
What am I missing here? Can someone help explain to me how I can get traffic flowing to my host? Or do I need to contact my datacenter, and ask them to stop being cheap, and allocate a /48 to me, or something?
/64
networks. You could get a/60
prefix that would give you a/64
transit network and 15/64
access networks. I'm guessing the data center guys really do not understand IPv6 or the fact it was designed to waste addressing, and they are trying to apply IPv4 address conservation that is detrimental to IPv6. – Ron Maupin Jun 28 '21 at 00:20/64
as a transit network, and you are supposed to use IPv6 Prefix Delegation to get the/32
from the data center router. – Ron Maupin Jun 28 '21 at 00:37/64
they gave me, and will work with them further on this. If they did just give me a basic/64
, I've asked them for a bigger prefix. – David W Jun 28 '21 at 00:50smallest
prefix (thanks, @RonMaupin for the clarification) they can give me is a /64. I'm considering going ahead and getting my own /40 direct from ARIN, but I'll see what they come up after I pointed out to them I'm unable to route anything right now. – David W Jun 28 '21 at 19:40/32
prefix from ARIN (RIPE gives out/29
). That means the datacenter has 4,294,967,296/64
networks (the same number of IPv6 networks as the total number of individual IPv4 addresses, including multicast and reserved). People just do not realize that IPv6 is designed to waste addressing, and trying to impose IPv4-style address conservation is detrimental. – Ron Maupin Jun 29 '21 at 00:29/64
networks, and if the world population is 21 billion in the year 2100 (a somewhat realistic number), every one of those 21 billion people could have 109,802,048 standard IPv6/64
networks, each network having 18,446,744,073,709,551,616 possible host addresses. – Ron Maupin Jun 29 '21 at 00:29