This repository contains an operator to manage the lifecycle of Dask clusters. It is a wrapper using Operator SDK around dask helm charts.
Prerequisites
- go
- make
- kustomize
make install deploy
TODO
The below code starts a cluster with 2 workers + 1 scheduler, and a jupyter notebook server.
k apply -f - << EOF
apiVersion: charts.dask.org/v1alpha1
kind: Dask
metadata:
name: dask-sample
spec:
# Default values copied from <project_dir>/helm-charts/dask/values.yaml
jupyter:
enabled: true
image:
repository: daskdev/dask-notebook
tag: 2021.12.0
name: jupyter
password: sha1:aae8550c0a44:9507d45e087d5ee481a5ce9f4f16f37a0867318c
rbac: true
replicas: 1
serviceType: ClusterIP
scheduler:
enabled: true
image:
repository: daskdev/dask
tag: 2021.12.0
name: scheduler
replicas: 1
servicePort: 8786
serviceType: ClusterIP
worker:
image:
dask_worker: dask-worker
repository: daskdev/dask
tag: 2021.12.0
name: worker
replicas: 2
EOF
To clean up,
kubectl delete dask dask-sample
For a comprehensive on the configs available, please refer to dask helm chart's value.yaml
To contribute, download operator-sdk.