Cluster management#

Overview#

A cluster provides the resources for Starburst Enterprise platform (SEP) to run queries against numerous data sources for the SEP users.

Clusters define the number of workers and their configuration for the JVM runtime system and others aspects, including the data sources. With Mission Control the creation and management of these clusters is available via a simple user interface.

The cluster management is suitable to create, update, and delete these clusters. Any user can access the cluster source management by clicking on the Clusters item in the left hand panel.

The cluster management view displays a grid of all existing clusters.

Press the New Cluster or New button on the screen to add a new cluster. The user creating a cluster is assigned as the owner of the cluster.

Each cluster is deployed to an individual Kubernetes namespace starting with the sep-ns- prefix, and a cluster can only accommodate a single namespace.

Clicking on a cluster name, or using the Edit link in the actions drop down, allows you to edit the cluster.

To use the cluster, you have to ensure that you have added the configured data sources, which you want to expose to the users of the particular cluster.

The cluster management view shows the following details for each cluster:

Name:

The user visible name of the cluster.

Actions drop down:

A list of actions for the cluster. Edit allows you to edit the cluster configuration. Clone allows you to make a copy of the cluster configuration. Delete allows you to delete the cluster configuration.

State:

The operational state of the cluster. The view includes a START button to start the cluster, or a STOP/RESTART button to change the state. In addition, a display for the number of workers is included. It includes functionality to scale the number of workers for a running cluster using + or - icons.

Data source:

A summary of the data sources configured for the cluster.

Failure Reason:

A short description of the failure reason if any of the cluster operations, such as starting the cluster, fails.

Cluster configuration#

You can create and edit a cluster with the help of a multiple-page form wizard. In create mode, you can navigate between steps with the Previous and Next buttons. In edit mode, you can click on any step name to navigate to it.

Kubernetes cluster configuration#

The following section describes the cluster configuration when deploying Mission Control to Kubernetes as detailed in the documentation about installing Mission Control.

Cluster summary#

In edit mode, on the right side there is a summary section for your cluster. URL links are only available for enabled features on a running clusters. This section shows the following details for each cluster:

State:

The operational state of the cluster.

Started at:

The date and time of the most recent cluster start.

Owner:

The user who created the cluster and therefore is considered the owner, detailed in the security section.

Created at:

The date and time of the initial cluster creation.

SEP Web UI:

The URL to the SEP WebUI.

The cluster forms focus on specific aspects about the cluster:

Data sources#

The data sources form allows you to configure the name of the cluster, and add and remove configured data sources.

SEP configuration#

The SEP configuration section allows you to configure many configuration values from SEP configuration.

Security configuration#

This section allows to you configure system level security using Ranger. See the Ranger plugin configuration for details about the configuration fields in this section.

General configuration#

The general configuration section can be used to configure selected extra settings like the Docker registry credentials or additional volumes.

Note

Additional volumes require a namespace. It needs to contain a PVC, secret or config map used by the SEP cluster, and can only accommodate one SEP cluster. If you want to deploy another cluster using additional volumes, a new namespace must be created before.