You can install dask-kubernetes with
conda, or by installing from source.
KubeCluster you may need to have
kubectl installed (official install guide).
HelmCluster you will need to have
helm installed (official install guide).
Pip can be used to install dask-kubernetes and its Python dependencies:
pip install dask-kubernetes --upgrade # Install everything from last released version
To install the latest version of dask-kubernetes from the conda-forge repository using conda:
conda install dask-kubernetes -c conda-forge
Install from Source¶
To install dask-kubernetes from source, clone the repository from github:
git clone https://github.com/dask/dask-kubernetes.git cd dask-kubernetes python setup.py install
pip locally if you want to install all dependencies as well:
pip install -e .
You can also install directly from git main branch:
pip install git+https://github.com/dask/dask-kubernetes
All Dask projects generally follow the NEP 29 deprecation policy for Python where each Python minor version is support ed for 42 months. Due to Python’s 12 month release cycle this ensures at least the current version and two previous versions are supported.
The Dask Kubernetes CI tests all PRs against all supported Python versions.
For Kubernetes we follow the yearly support KEP. Due to the 4-6 month release cycle this also ensures that at least the current and two previous versions are supported.
The Dask Kubernetes CI tests all PRs against all supported Kubernetes versions.
To keep the CI matrix smaller we test all Kubernetes versions against the latest Python, and all Python versions against the latest Kubernetes. We do not test older versions of Python and Kubernetes together. See dask/dask-kubernetes#559 for more information.