Assigning Remote Engines to clusters - Cloud

Talend Remote Engine User Guide for Linux

author
Talend Documentation Team
EnrichVersion
Cloud
EnrichProdName
Talend Cloud
task
Design and Development
Installation and Upgrade
EnrichPlatform
Talend Remote Engine

Before you begin

It is recommended to only add Remote Engines with the same version to a cluster.

Procedure

  1. Go to the Engines page.
  2. Click on the name of cluster you to which you want to add Remote Engines.
  3. Click the icon.

    Example

  4. Select the Remote Engines to include in the cluster.

    You can only select Remote Engines from the same environment where your cluster is.

    You can select Remote Engines from any workspace on that environment. However, if the workspaces are different, the selected Remote Engines will be placed into the cluster's workspace.

    Note: When deploying Routes and data services on a cluster, all Remote Engines assigned to it must be of the same version (v2.8.4 or above).
  5. If tasks are already running on the Remote Engines you selected, click RESCHEDULE on the dialog box that appears to run the tasks on your cluster.

    Example

Results

When choosing to run a Job task on the cluster, it is executed in a round-robin manner on the engines belonging to the cluster. The task is executed on the first available Remote Engine, then the next and so on, while putting the ones the task was executed on to the end of the list. If a Remote Engine is not available, the task is executed on the next available engine.

When choosing to run a data service or Route task on the cluster, you must define a deployment strategy.
Parallel The tasks are deployed simultaneously to all the Remote Engines of the cluster. Before deploying a new version of a task, you will need to undeploy the currently running version, which will result in service interruption.
Rolling If the Remote Engine cluster contains more than one Remote Engine, the tasks are deployed one by one to the Remote Engines of the cluster. This means that for a short time, different Remote Engines will be running different versions of the task, therefore you can avoid service interruption.