Deploying a custom image
In case you might want to deploy your own docker image of on our more of the servers the easiest way to do this is to deploy to a custom image to a public dockerhub. Currently the Helm files assume a public dockerhub repository. If you want it be private an key for authenticating with Dockerhub, or another registry should be added.
Most likely the custom image will be a version of the Frontend. So for instance in your values yaml overwrite the name of the image:
frontend:
# Deployment container is what is eventually run in the pod
deploymentContainer:
name: frontend
# Docker image for this pod
image: dockerhubname/imagename:tag
Run upgrade command:
helm upgrade -f custom-values.yaml openstad . -n openstad
Every repository has a travis.yml file that would work automatically if a travis account was made dockerhub images are pushed. Both Dockerhub and Travis are free for open source project.
For instance the frontend has a .travis.yml file and de docker_publish script: https://github.com/Amsterdam/openstad-frontend/tree/development
The publish script cert environment variables, these can be saved in repository settings in the Travis dashboard:
- DOCKER_USERNAME
- DOCKER_PASSWORD
It's possible, we do it for staging and release, to autodeploy to a Kubernetes cluster. See the docker_deploy file in the repository for this.
For deploy to work, the above variables are necessary and needed to add the kubernetes config file as a variable:
- KUBE_CONFIG
This is the kubeconfig generated encoded to a base64 string. For instance can be generated like this:
cat ${HOME}/.kube/config | base64 | pbcopyLast modified 1yr ago