Reach's pipeline is deployed in production using Argo. These files are for local runs and development.
To run this pipeline locally, you'll need:
- Docker
- Minikube
- Python >= 3.6
- The Argo cli (recommended but optionnal)
If it is the first time you use minikube with your AWS account, please configure your ECR credentials:
minikube addons enable registry-creds
minikube addons configure registry-creds
To build the required images, go to the root folder and run the following:
make docker-build
To install Argo to your selected cluster (this will install Argo to a namespace argo, so make sure it's available before running these commands or change it beforehand):
kubectl apply -f argo/00-namespace.yaml
kubectl apply -f argo/argo.yaml
kubectl apply -f argo/elasticsearch.yaml
kubectl apply -f argo/psqlinit.yaml
kubectl apply -f argo/postgres.yamls
You can then run your workflows as follows:
# this is the example workflow for WHO IRIS
argo submit -n argo argo/reach-who.yaml
Reach's web application only relies on Postgresql. To expose it and make it usable locally (or within the docker-compose local deployment), run:
kubectl port-forward -n argo postgres-0 5432:5432