.. _deployment_cloud: ================ Cloud deployment ================ Prerequisites ------------- Currently, we support deploying Xorbits on an existing Amazon EKS cluster. Install Xorbits on the machine for Amazon EKS cluster management. Refer to :ref:`installation document `. Amazon EKS ---------- Firstly, make sure your EKS cluster can access `our Dockerhub `_ to pull the Xorbits image. Secondly, install the `AWS Load Balancer Controller `_. Then, deploy Xorbits cluster, for example: .. code-block:: python from kubernetes import config from xorbits.deploy.kubernetes import new_cluster cluster = new_cluster(config.new_client_from_config(), worker_cpu=1, worker_mem='4g') Note that the option ``cluster_type`` of the function ``new_cluster`` has default value ``auto``, which means that Xorbits will detect the ``kubectl`` context automatically. Please make sure ``kubectl`` is using the correct EKS context. You should be able to see ``Xorbits endpoint http://:80 is ready!`` soon, and you can access the web UI of your Xorbits cluster using the endpoint. Refer :ref:`Kubernetes deployment ` to deploy Xorbits with your own image. Refer :ref:`Install Python Packages ` to install additional python packages for the Xorbits supervisors and workers. ``new_cluster`` api refers to :meth:`xorbits.deploy.kubernetes.client.new_cluster`. To verify the cluster: .. code-block:: python import xorbits.numpy as np a = np.ones((100, 100), chunk_size=30) * 2 * 1 + 1 b = np.ones((100, 100), chunk_size=20) * 2 * 1 + 1 c = (a * b * 2 + 1).sum() print(c) External Storage ---------------- Great news! Xorbits starts to support separation of storage and compute. Start to enjoy better scalability, flexibility, and performance it brings. Check out the last section ``JuiceFS on Kubernetes`` on the page :ref:`Kubernetes Deployment `.