That's good news.
Can you clients reach WAN via IPv6 now?
No... still more problems.
Okay, so here's the steps:
1. Initially, the router itself did receive a DHCP assignment and prefix delegation from the upstream gateway on the WAN interface, but it could not reach the internet
2. It seemed the router was unable to reach the internet because there was no default route. When I added:
ipv6/route/add gateway=2001:df6:5880:1000::1 dst-address=::/0
add a static default route to the upstream gateway, the router was able to reach the internet, but LAN clients couldn't.
3. LAN clients couldn't reach the internet, apparently because the gateway address issued to the LAN clients was a link-local address, not the LAN interface's actual given address
. The router seemed to have no link-local addresses on any interfaces though.
4. I disabled/reenabled ipv6, and the link-local addresses reappeared for all interfaces. NOTE: it is THE SAME LINK-LOCAL ADDRESS ON ALL INTERFACES... should each interface have its own address?
5. LAN clients can now reach the router, but they can still not reach the internet...
6. The router can no longer reach the internet "no route to host" again... I checked the default route again, and noticed that disable/reenable ipv6 caused my static `::/0` route to disappear.
7. I re-add the default route (same command as above), and the router can reach the internet again.
8. LAN clients can still not reach the internet; but instead of 'no route to host', they now get 'request timed out' when trying to ping the upstream router (the very next hop) or google DNS, which the router can ping itself.
Tracing route to dns.google [2001:4860:4860::8888]
over a maximum of 30 hops:
1 <1 ms <1 ms <1 ms 2001:df6:5880:10fb:2ec8:1bff:feff:9ef8
2 * * * Request timed out.
3 * * * Request timed out.
4 * * * Request timed out.
5 * * * Request timed out.
6 * * * Request timed out.
I'm really feeling like like something has gone terribly wrong somewhere.
Should all interfaces have the same link-local address, or is that a configuration fault?
[admin@cw-lot32] > ipv6/address/print
Flags: D - DYNAMIC; G, L - LINK-LOCAL
Columns: ADDRESS, FROM-POOL, INTERFACE, ADVERTISE
# ADDRESS FROM-POOL INTERFACE ADVERTISE
0 G 2001:df6:5880:10fb:2ec8:1bff:feff:9ef8/64 isp_pool bridge yes
1 DG 2001:df6:5880:1000::fc/64 bridge.30 no
2 DL fe80::2ec8:1bff:feff:9ef8/64 bridge.30 no
3 DL fe80::2ec8:1bff:feff:9ef8/64 bridge.102 no
4 DL fe80::2ec8:1bff:feff:9ef8/64 bridge.2 no
5 DL fe80::2ec8:1bff:feff:9ef8/64 bridge no
6 DL fe80::2ec8:1bff:feff:9ef8/64 bridge.666 no
7 DL fe80::2ec8:1bff:feff:9ef8/64 bridge.3 no
[admin@cw-lot32] > ipv6/route/print
Flags: D - DYNAMIC; A - ACTIVE; c, s, y - COPY
Columns: DST-ADDRESS, GATEWAY, DISTANCE
# DST-ADDRESS GATEWAY DISTANCE
0 As ::/0 2001:df6:5880:1000::1 1
DAc 2001:df6:5880:1000::/64 bridge.30 0
DAc 2001:df6:5880:10fb::/64 bridge 0
DAc fe80::%bridge/64 bridge 0
DAc fe80::%bridge.2/64 bridge.2 0
DAc fe80::%bridge.30/64 bridge.30 0
DAc fe80::%bridge.666/64 bridge.666 0
DAc fe80::%bridge.102/64 bridge.102 0
DAc fe80::%bridge.3/64 bridge.3 0
Can I influence the router to include proper interface addresses as the gateway address, rather than the link local address? I don't want link-local addresses in the RA's to LAN clients, they make the whole thing really hard to understand and debug.