[Kea-users] Believe my config isn't working well with customer devices

[Chris Gross]

I have a question about some configuration to see if I'm making my config too simple and thus broken, or if I may have something misconfigured on my network.

Currently I build out a block for a /48 of WAN blocks, and anything in that /48 is allowed to assign a /56 from the /37. Then on my Cisco device, I have a general-prefix named DHCP set for the /48, and assign a /64 from that to each vlan I need a DHCPv6 space.

           {
            "subnet": "2001:db8:2001::/48",
            "pools": [ { "pool": "2001:db8:2001::/48" } ],
            "pd-pools": [
                {
                    "prefix": "2001:db8:2800::",
                    "prefix-len": 37,
                    "delegated-len": 56
                }
            ]
        },

ipv6 general-prefix DHCP 2001:db8:2001::/48

interface VlanXXYY
vrf forwarding CGNAT
ip address 100.XX.XX.1 255.255.252.0
ip helper-address global 10.0.0.3
ip helper-address global 10.0.0.4
ip flow ingress
load-interval 30
ipv6 address FE80::1 link-local
ipv6 address DHCP ::5:0:0:0:X/64
ipv6 enable
ipv6 nd managed-config-flag
ipv6 nd other-config-flag
ipv6 nd router-preference High
ipv6 dhcp relay destination 2001:db8::1
ipv6 dhcp relay destination 2001:db8::2
ipv6 verify unicast source reachable-via any
end

I see plenty of routes taken in my table, but very few IPv6 neighbors and seeing very little IPv6 traffic traversing, especially in comparison to Netflix and other for sure IPv6 sources.

Do I need to expand the /48 "subnet" into a list of /64 pools to assign from to work more efficiently? Missing a flag of some sort in my configuration to indicate to customer routers a smoother experience?

Chris Gross
Network Architect
cgross at ninestarconnect.com<mailto:cgross at ninestarconnect.com>
www.ninestarconnect.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/kea-users/attachments/20200409/2a001267/attachment.htm>