- image is now based on the official mono base image which is based on Debian
- image is available for amd64, arm32v5 and aarch64
- Images are built on Docker Hub for various architectures (see
hooks/.config
) - Build is triggered by git a push to https://github.com/rmoriz/openra-dockerfile
- Manual build (without multiarch):
./hooks/build
(works on Mac) - Manual build (with multiarch on Debian/Ubuntu Linux):
apt-get install qemu-user-static cp /usr/bin/qemu-{aarch64,arm,x86_64}-static . ./hooks/build
$ docker run -d -p 1234:1234 \
-e Name="DOCKER SERVER" \
-e Mod=ra \
-e ListenPort="1234" \
--name openra \
rmoriz/openra
see https://github.com/OpenRA/OpenRA/blob/bleed/launch-dedicated.sh for all options.
kubectl apply -f k8s.yml
If you have a l7 load balancing nginx with tcp services enabled you can add this to the configmap, the format is: port for key, value is namespace/servicename:port
1234 -> openra/openra-service:1234
You can also add a nodeport to the service descriptor:
type: NodePort
ports:
- port: <1234>
# nodePort: <31514>
If the nodePort is not specified, k8s will assign you one but it will be random.
The latest release is available for amd64/x86_64 and arm32v5. Usually docker should automatically download the right image.
docker pull rmoriz/openra:<TAG>
should just work (tm)
Naming schema:
rmoriz/openra:<TAG> = Manifest with all versions
rmoriz/openra:<TAG>-amd64 = Image for x86
rmoriz/openra:<TAG>-arm = Image for arm32v5 (Raspberry Pi 1, Zero or better)
rmoriz/openra:<TAG>-aarch64 = Image for aarch64/ARMv8 (Raspberry Pi 3b, 3b+ with 64bit distro, does not work on stock Raspbian yet)
Seems like an example server with RA-mod needs at least 500MB of memory, maybe even more.
Copyright © 2019 Roland Moriz, see LICENSE.txt