Starting and Stopping the Spark Cluster

Owned by Kevin Wu
Jan 03, 2024

Note!

As a prerequisite, the scripts must be prepared according to Preparing and Creating Scripts for KPI Management.

Starting the Spark Cluster

The Spark cluster is started automatically when you start the Spark service. For information about how to configure and start this service, see KPI Management - External Software plus the subsections thereof.

$ start_master_workers.sh

For further information about the Spark command, see /wiki/spaces/temp/pages/23280759.

Stopping the Spark Cluster

$ stop.sh

For further information about the Spark command, see /wiki/spaces/temp/pages/23280759

Note!

The cluster stop command does not automatically trigger the output of the currently processed KPIs.To flush all data, you must start a new period for your KPIs. For further information, see KDR.

Alternatively, you can manually flush all the pending KPIs to the output topic, by using the spark by using the flush scripts, which triggers an output of previously unclosed periods. Example:

flush.sh kpiapp