I've asked this question on the Docker support forum (no email
available!) but have seen no answers. If there are other better places
to ask, please let me know.
I have a Fedora 23 host (to be upgraded to F25 or F26 in the near
future). It has two networks and I want to run a Docker container that
can participate in both of them. The host defines these two networks
(and a gateway on a third network),
$ route -n
Destination Gateway Genmask ... Iface
default w.x.y.254 0.0.0.0 ... eth0
a.b.c.0 0.0.0.0 255.255.255.0 ... eth2
e.f.g.0 0.0.0.0 255.255.255.0 ... eth3
So far for Docker, I have,
(1) $ docker network create --subnet a.b.c.0/24 --gateway a.b.c.254 eth2
(2) $ docker network create --subnet e.f.g.0/24 --gateway e.f.g.254 eth3
(3) $ docker create --network=eth2 --ip=a.b.c.1 container program
(4) $ docker network connect --ip=e.f.g.1 eth3 container
(5) $ docker start container
# (With a few more arguments in real life.)
This gets me most of what I want except the ability to actually
participate in the host's networks. Traffic doesn't leave the container.
I understand I must somehow map the container's networks to the host's
networks. But I'm having trouble learning how to do that.
I can map the first network, I think. Instead of line (3) above, I try,
(6) $ docker create --network=eth2 --ip=a.b.c.1 \
-p a.b.c.d:1-65535:1-65535/tcp -p a.b.c.d:1-65535:1-65535/udp \
container program
But I don't see how to map the second network since the -p option is not
available for "docker network connect" (or for "docker start"). Of
course, using -p might not be the proper solution, anyway. Is it
possible to do what I want?
I'm using docker-engine-1.12.6-1.fc23.x86_64 and associated packages
from Docker, not Fedora's own Docker packages, because the latest
available from Fedora for F23 is Docker 1.10.
--
Dave Close
"Age is a very high price to pay for maturity." -- Tom Stoppard
_______________________________________________
users mailing list -- users@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe send an email to users-leave@xxxxxxxxxxxxxxxxxxxxxxx
