@ariadne@social.treehouse.systems Heck how does your own router even know how to route delegated prefixes to other routers, if it doesn't have a full routing table for them? Just bouncing off their headend every time? If you ping another customer's SLAAC router IP from a PD subnet IP, does the request go directly within the subnet and the reply get bounced through the headend?!?!
@lina@vt.social
# ip -6 route show
anycast 2604:5500:706b:: dev eth1 proto kernel metric 0 pref medium
2604:5500:706b::/64 dev eth1 proto ra metric 1005 expires 2591846sec mtu 1500 pref medium
anycast 2604:5500:706b:3e00:: dev eth0 proto kernel metric 0 pref medium
2604:5500:706b:3e00::/64 dev eth0 proto dhcp metric 1003 expires 21474836sec pref medium
unreachable 2604:5500:706b:3e00::/56 dev lo proto dhcp metric 1001 pref medium
@lina@vt.social
and of course default via the link-local address on the headend:
default via fe80::629c:9fff:feb4:2ca8 dev eth1 proto ra metric 1005 expires 1646sec mtu 1500 pref medium
@ariadne@social.treehouse.systems This is so cursed...
@lina@vt.social @ariadne@social.treehouse.systems incredibly cursed yet sadly not uncommon with busted US ISPs :/
@astraleureka@social.treehouse.systems @lina@vt.social yeah but this DHCPv6-PD issue is just next-level busted
@ariadne@social.treehouse.systems @astraleureka@social.treehouse.systems @lina@vt.social Sadly not rare.
my last dhcpv6-pd issue (comcast business):
1) you must request your static /56 as many separate /59 delegations, it will refuse to delegate anything larger in one go
2) you must enable IPv4 DHCP on the CPE. Doesn't matter that your router is on a static ip, or that you will not actually ever issue a DHCPDISCOVER. If the DHCPv4 service isn't running, it'll send you DHCPv6 messages claiming to have delegated your prefix but not actually update the routing table on the CPE and your prefix won't route anywhere
@azonenberg@ioc.exchange @ariadne@social.treehouse.systems @astraleureka@social.treehouse.systems @lina@vt.social it could be WAY worse:
when I got my fiber connection, I found for some reason it worked on OpenWRT but not my Debian box. After some digging, it turned out all DHCPv6 packets were 1514 byte Ethernet frames, regardless of IPv6/UDP content len.
Those extra bytes were a bug on the last-hop Cisco switches' DHCPv6 relay/inspection. Random snippets out of the switch's memory.
Including configs.
With other customer's names and line IDs (port descriptions)
Asahi Lina (ææ„ăȘă) // nullptr::live
Ariadne Conill đ°
:therian:
