Docker SkinnyWMS

Attention: My pull request was merged to the ECMWF SkinnyWMS repository at https://github.com/ecmwf/skinnywms, so now an official docker image ecmwf/skinnywms is available on dockerhub!

Please use the official image!

docker run --rm -i -t -p 5000:5000 ecmwf/skinnywms

Git

Docker

skinnywms packaged to run in a docker container with a simple demo application.

Run pre-built image with Docker

The pre-built image eduardrosert/skinnywms is available on Dockerhub. If you already have Docker running on your machine, just do the following to run the image.

Run the image on your machine

Run the demo applicaton and publish port 5000. The --rm switch makes sure that docker leaves no temporary files behind, when you stop the image:

docker run --rm -i -t -p 5000:5000 eduardrosert/skinnywms

Now you can type http://localhost:5000 on the machine running docker to access the demo application.

Run interactive shell

Run interactive shell /bin/bash in docker image eduardrosert/skinnywms or any other command for that matter overriding the default CMD instruction from the docker image:

docker run --rm -i -t -p 5000:5000 eduardrosert/skinnywms /bin/bash

Installation instructions

Follow these steps to manually create the docker image and make it available locally.

Automatic image creation using make

#go to docker-skinnywms folder and run
make

#check if creation was successful
docker images

REPOSITORY                                TAG                 IMAGE ID            CREATED             SIZE
eduardrosert/skinnywms                    latest              eb4953292e7f        1 minute ago        727MB
...

Manually run image using kubernetes

If you want use a local image and don't want to pull it from dockerhub, make sure you tag the image first, otherwise by default kubernetes will try to pull the most recent version from a public repo.

docker tag eduardrosert/skinnywms:latest eduardrosert/skinnywms:localOnly

Create deployment

kubectl create deployment skinnywms --image eduardrosert/skinnywms:localOnly

Check pod status

kubectl get pods

result should look something like this:

NAME                         READY   STATUS    RESTARTS   AGE
skinnywms-756dffddd4-2xrgk   1/1     Running   0          18s

Expose deployment:

kubectl expose deployment skinnywms --type=LoadBalancer --port=5000

or if you are using a custom kubernetes cluster such as minikube or kubeadm

kubectl expose deployment skinnywms --type=NodePort --port=5000

List services

kubectl get service

you should see the service:

NAME         TYPE           CLUSTER-IP       EXTERNAL-IP   PORT(S)          AGE
skinnywms    LoadBalancer   10.102.228.234   <pending>     5000:31508/TCP   10s
...

Make service accessable from outside

minikube service skinnywms

Get service url

minikube service skinnywms --url

example output:

http://192.168.99.101:31751

Access the demo application directly (starts a browser window)

minikube service skinnywms

In this example the GetCapabilities document should then be found at http://192.168.99.101:31751/wms/?request=GetCapabilities