On Fri, Feb 23, 2018 at 10:45:24AM -0500, Stephen P. Molnar wrote:
> root@AbNormal:/home/comp# ip netns exec test ip a l
> 3: net0@if2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state
> UP group default qlen 1000
> link/ether be:80:71:d1:8a:96 brd ff:ff:ff:ff:ff:ff link-netnsid 0
> inet6 2600:1700:4280:3690:bc80:71ff:fed1:8a96/64 scope global mngtmpaddr
> valid_lft 1209450sec preferred_lft 1209450sec
> inet6 fe80::bc80:71ff:fed1:8a96/64 scope link
> valid_lft forever preferred_lft forever
So, I have a good news, and a bad news.
Good news being - I honestly don't know how you were able to achieve
*that* IPv6 configuration in a primary network namespace, but in this
separate network namespace things look reasonable.
You have one RA-provided IPv6 address, which is normal if one disables
IPv6 privacy extensions (they are disabled by default, btw).
> root@AbNormal:/home/comp# ip netns exec test ip -6 ro l
> 2600:1700:4280:3690::/64 dev net0 proto kernel metric 256 expires
> 1209439sec pref medium
> fe80::/64 dev net0 proto kernel metric 256 pref medium
> default via fe80::3e04:61ff:feb3:3c20 dev net0 proto ra metric 1024 expires
> 1639sec hoplimit 64 pref medium
And you have perfectly normal IPv6 routing table, with RA-provided
> root@AbNormal:/home/comp# ip netns exec test traceroute -n
> traceroute to 2a02:16a8:dc41:100::233 (2a02:16a8:dc41:100::233), 30 hops
> max, 80 byte packets
> 1 * * *
Which brings me to the bad news.
Whatever router you're using refuses forwarding your IPv6 packets.
It does not matter if it drops the packets, or sends your host some
"refused" messages via SNMP - the thing fails to perform its primary
I deliberately stay clear from SOHO routers, regardless of whoever
produced them, so I cannot help you here. In fact, I choose mine with
exactly one quality in mind - an ability to run Debian. Which I
installed on it the moment they delivered me the thing.
Best advice I can give - get yourself something that can be flashed with
Until then - disable IPv6 on your router altogether, it's not going to