i'm having serious problems enabling IPv6 in docker.
Environment
- The host is running Debian Jessie.
- It's a virtual Server (KVM).
- eth0 has a statically configured address like w:x:y:z::1 in a network like w:x.y:z::/64, which is assigned to my by my hosting company.
- My host is capable of using IPv6 without any problem: Pinging the outside world works, a website running on a container (Port 80 bound to host:80) is accessible via ipv6.
Problem
I cannot however access the outside world from within the containers! My docker0 bridge does NOT have a IPv6 address after restarting docker with the parameters below. There's no route and no gateway either (doesn't makes sense without an ipv6 address).
My Docker setup: Docker is started with these parameters in DOCKER_OPTS
DOCKER_OPTS="--dns 8.8.8.8 --dns 8.8.4.4 --ipv6 --fixed-cidr-v6=w:x:y:z:a::/80"
Some ipv6 host configuration parameters:
net.ipv6.conf.all.forwarding = 1
net.ipv6.conf.default.forwarding = 1
here's One of the networks i created myself:
root@wopr:~# docker network inspect wopr6
[
{
"Name": "wopr6",
"Id": "ddc192d4af2a8edc809975e84cf3e4cb82c24d4cfe970dd8e3fc7d6ff31e20ee",
"Scope": "local",
"Driver": "bridge",
"EnableIPv6": true,
"IPAM": {
"Driver": "default",
"Options": {},
"Config": [
{
"Subnet": "172.23.0.0/16",
"Gateway": "172.23.0.1/16"
},
{
"Subnet": "w:x:y:z:a:0:0:0/80",
"Gateway": "w:x:y:z:a::1"
}
]
},
"Internal": false,
"Containers": {
"dff30ab1496a4c3689ad6da0837fdb6cf7ea1a5b32312116214313b5b14ed07e": {
"Name": "happy_varahamihira",
"EndpointID": "8cd4ed4b91d8421171ec8cc771bbe7b7d81f05dc9f4679f20c642c2e828ec475",
"MacAddress": "02:42:ac:17:00:02",
"IPv4Address": "172.23.0.2/16",
"IPv6Address": "w:x:y:z:a::2/80"
}
},
"Options": {},
"Labels": {}
}
]
Here's some information from inside the container, which is mentioned above:
Addresses
root@dff30ab1496a:/# ip -6 a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
332: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500
inet6 2a03:4000:6:2158:a::2/80 scope global nodad
valid_lft forever preferred_lft forever
inet6 fe80::42:acff:fe17:2/64 scope link
valid_lft forever preferred_lft forever
Routes
root@dff30ab1496a:/# ip -6 r
2a03:4000:6:2158:a::/80 dev eth0 proto kernel metric 256
fe80::/64 dev eth0 proto kernel metric 256
default via 2a03:4000:6:2158:a::1 dev eth0 metric 1024
Ping
PING ipv6.l.google.com (2a00:1450:4001:811::200e): 56 data bytes, id 0x0011 = 17
--- ipv6.l.google.com ping statistics ---
1 packets transmitted, 0 packets received, 100% packet loss
What am i missing?
So after waiting two weeks for an answer and researching another few hours after opening up a bounty i found the solution.
Set up a new IPv6 enabled network and assign a subnet available to me (a /80 of my /64)
Now start a container and connect it to the new network. Find out it's IP address. Let's say it's w:x:y:z:aaaa::5 in this example.
Enable proxy_ndp
You may also configure this setting via /etc/sysctl.conf, to make it persistent.
Add proxy to make my host (IPv6 enable) to respond to Neighbour Sollicitation messages from my router (like: "hey, who's hosting w:x:y:z:aaaa::5?") with Neighbour Advertisement messages ("that would be me!").
ndppd may help you to automatically advertize any hosts on your network.
Bam, that's it.