WEKA Operator deployments

Deploy WEKA on Kubernetes with the WEKA Operator. Choose the deployment model that fits your environment, verify version support, and continue to the right workflow.

WEKA Operator capabilities

The WEKA Operator automates the deployment and lifecycle management of WEKA storage on Kubernetes. It manages the full stack through Kubernetes-native custom resources, removing the need for manual coordination across cluster components.

The Operator handles the following:

Resource provisioning: Deploys and manages WekaCluster and WekaClient custom resources as Kubernetes-native objects.
Driver lifecycle: Distributes pre-built drivers and supports local driver build workflows for custom or restricted environments.
CSI management: Embeds and manages the WEKA CSI Plugin lifecycle as part of the WekaClient resource.
Observability and monitoring: Exposes WEKA health and performance metrics to Kubernetes monitoring stacks.
Protocol gateways: Manages protocol containers, including S3, NFS, SMB, and audit, from the WekaCluster resource.
Drive sharing and composable backends: Supports drive sharing configuration, drive type ratios, and capacity-based backend layouts.
Upgrade management: Supports rolling, manual, and all-at-once upgrade policies for client pods.
Cluster maintenance controls: Supports CSI migration, cluster pause and resume, and cluster deletion cancellation.

The WEKA Operator is updated frequently. For the latest feature additions and changes, always refer to the release notes for your target Operator version before deployment or upgrade. For example: WEKA Operator 1.14.0 release notes.

Deployment models

Choose the model that matches your infrastructure and operating boundary.

Full deployment on Kubernetes: Run WEKA backend and client processes in Kubernetes. Use this model when Kubernetes hosts both the WEKA data plane and application workloads.

Client-only deployment: Run WEKA clients in Kubernetes and connect them to a WEKA cluster running outside Kubernetes. Use this model when the storage cluster already runs on dedicated servers.

Managed Kubernetes services: Run supported WEKA components on Amazon EKS, Oracle OKE, Google GKE, or Azure AKS. Platform support and deployment limits vary by provider and instance type.