You can deploy flowmium on your local machine for testing and for production
-
Install flowctl
cargo install flowmium -
Run a registry and kubernetes on your machine as docker containers
docker-compose up -
Deploy flowmium onto the cluster
KUBECONFIG=./kubeconfig.yaml kubectl apply -f kubernetes.yaml -
Watch Kubernetes pods
KUBECONFIG=./kubeconfig.yaml watch kubectl get pods -
Expose flowmium server
KUBECONFIG=./kubeconfig.yaml kubectl port-forward svc/flowmium-server-service 8080 -
Watch for flows status
watch flowctl list
You can use kubernetes.yaml in this example folder for deploying flowmium but replace the postgres and minio deployments with other helm charts. Flowmium is also not designed to be run as multiple instances or replicas. You may also want to configure the server by setting following environment variables
| Name | Description | Example |
|---|---|---|
| FLOWMIUM_POSTGRES_URL | URL to postgres database | postgres://flowmium:flowmium@localhost/flowmium |
| FLOWMIUM_STORE_URL | URL to s3 compatible storage like MinIO | http://172.16.238.4:9000 |
| FLOWMIUM_TASK_STORE_URL | URL to s3 from within the cluster, this will be the same as FLOWMIUM_STORE_URL for most cases, this would be diffrent if s3 and the server are running outside the cluster |
http://172.16.238.4:9000 |
| FLOWMIUM_BUCKET_NAME | Name of the bucket to store artefact in | flowmium-test |
| FLOWMIUM_ACCESS_KEY | Access key for s3 | minio |
| FLOWMIUM_SECRET_KEY | Secret key for s3 | password |
| FLOWMIUM_INIT_CONTAINER_IMAGE | Image to use for the init container | docker.io/shnoo28/flowmium:latest |
| FLOWMIUM_NAMESPACE | Namespace to spawn or deploy jobs in | default |
| KUBECONFIG | Path to kubeconfig, not required if a Kubernetes service account is attached | ./kubeconfig.yaml |