Managing Legacy Deployments

Applies to: IBM StreamSets as a Service

After you create a legacy deployment, you can start, stop, scale, or delete the deployment. You must stop a deployment to specify a different Provisioning Agent for the deployment, specify a different Data Collector Docker image to deploy, or assign additional labels to the deployment.

Most actions that you complete on a legacy deployment instruct the assigned Provisioning Agent to complete a corresponding action within Kubernetes.

The following table lists legacy deployment actions that you can complete from the Legacy Deployments view, and any corresponding actions that the Provisioning Agent completes in Kubernetes:

Legacy deployment action Action completed in Kubernetes
Create a deployment None.

Creating a deployment simply creates the deployment configuration and saves the configuration in Control Hub.
Assign labels to a deployment None.

Assigning labels to an inactive deployment simply modifies the deployment configuration saved in Control Hub.

Start a deployment The Provisioning Agent deploys the Data Collector containers to Kubernetes pods, starts each Data Collector container, and registers each Data Collector container with Control Hub.
Stop a deployment The Provisioning Agent stops the existing Data Collector containers and unregisters the Data Collector containers from Control Hub.
Manually scale a deployment When you scale up, the Provisioning Agent deploys, starts, and registers additional Data Collector containers in Kubernetes pods.
When you scale down, the Provisioning Agent stops the existing Data Collector containers and unregisters the Data Collector containers from Control Hub.
Note: If you associated a Kubernetes Horizontal Pod Autoscaler with the deployment, then Kubernetes handles the automatic scaling of the deployment.
Delete a deployment None.

Deleting an inactive deployment simply deletes the deployment configuration from Control Hub.

Manually Scale the Legacy Deployment

When you provision Data Collectors, you benefit from all of the features that Kubernetes offers - including manually scaling Data Collector containers during times of peak performance. You simply modify a legacy deployment in Control Hub to scale the number of Data Collector instances up or down. Then the Provisioning Agent and Kubernetes automatically take care of provisioning additional instances or removing instances that are no longer needed.

Note: If you associated a Kubernetes Horizontal Pod Autoscaler with the deployment, then Kubernetes handles the automatic scaling of the deployment.

When you scale a deployment up, the Provisioning Agent deploys, starts, and registers additional Data Collector containers in Kubernetes pods.

When you scale a deployment down, the Provisioning Agent stops the existing Data Collector containers and unregisters the Data Collector containers from Control Hub.

Note: Control Hub automatically scales out pipeline processing for an active job only when the number of pipeline instances for the job is set to -1. When the number of pipeline instances is set to any other value, you must synchronize the active job to start additional pipeline instances on the newly provisioned Data Collector containers.
  1. On the Legacy Kubernetes > Legacy Deployments view, click the active deployment to display its details.
  2. Increase or decrease the Number of Data Collector Instances.
  3. Click Scale.
    It can take the Provisioning Agent up to a minute to provision or remove Data Collector containers.
  4. To verify that the Data Collector containers were successfully scaled up or down, click Set Up > Engines in the Navigation panel.

Legacy Deployment Status

When you view the list of deployments in the Legacy Kubernetes > Legacy Deployments view, you can view the legacy deployment status. The deployment status is color-coded as an easy visual indicator of which deployments need your attention. A red status indicates that an error has occurred that you must resolve. An orange status indicates that a warning has occurred that you should look into. A green status indicates that all is well.

The following table describes each deployment status:

Legacy Deployment Status Description
Deployment is inactive. No provisioned Data Collector container instances are running for this deployment.

You can start, edit, and delete inactive deployments. Editing includes changing the Provisioning Agent that manages the deployment, changing the Data Collector container image to deploy, and changing the labels assigned to the deployment.
The Provisioning Agent is in the process of starting the deployment.

You cannot perform actions on activating deployments.
Deployment is active. One or more provisioned Data Collector container instances are running for the deployment.

You can stop and scale active deployments. Scaling includes increasing and decreasing the number of provisioned Data Collector container instances for the deployment.
The Provisioning Agent encountered an error during the activation process and was not able to start the deployment.

You cannot perform actions on deployments with an activating error status until you acknowledge the error message. To acknowledge deployment errors, click the row listing the deployment to display the deployment details. The details list the error message. Review the message, taking action as needed, and then click the Acknowledge Error icon: .
The Provisioning Agent is in the process of stopping the deployment.

You cannot perform actions on deactivating deployments.
Deployment is inactive and has an error that you must acknowledge.

This status can occur when you stop a deployment and the Provisioning Agent reports an error while attempting to stop the existing Data Collector containers.

You cannot perform actions on deployments with an inactive_error status until you acknowledge the error message. To acknowledge deployment errors, click the row listing the deployment with the inactive error to display the deployment details. The details list the error message. Review the message, taking action as needed, and then click the Acknowledge Error icon: .