Skip to content

Workflow Configuration

The Altair Analytics Workbench™ workflows have the requirement that the paths used are file shares mounted to the same place in all nodes of Altair SLC Hub.

Before you start

Goal

  • Run Altair Analytics Workbench workflows with link sessions on Hub.

Prerequisites

When a workflow is run, parts of it can be run in parallel and this could potentially be across multiple different Altair SLC Hub worker nodes. Since the intermediate results need to be available to all workflow nodes regardless of which Altair SLC Hub worker node the workflow node is run on, the directory the intermediate files are stored on needs to be on a shared filesystem mounted equally into each Altair SLC Hub worker nodes.

The user should have write access to the configured directory. This should also be unique to each user to prevent users interfering with each other. It is therefore highly recommended to use one of the substitution variables

Execution Profile Configuration

In the Altair SLC Hub portal, launch Hub Administration and select Execution Profiles. Select the execution profile to change, and set the Intermediate files directory field to a path like the example below and click Save.

/home/${userId}/workflowIntermediateFiles/

C:/Users/${userId}/workflowIntermediateFiles/

Legacy Configuration

Older Altair Analytics Workbench versions do not use the value per execution profile, they can be configured by setting the path in the workbench_custom.yaml file.

workbench:
  workflow:
    intermediateFilesDir: /home/${userId}/workflowIntermediateFiles/

workbench:
  workflow:
    intermediateFilesDir:  C:/Users/${userId}/workflowIntermediateFiles/

Verify

Verify that workflows can be run

To verify, launch Altair Analytics Workbench and establish a connection to the Hub. Select the Workflow perspective, and open the Workflow Link Explorer view. Right click on the Hub connection and select properties. Select to the Workflow Properties section and verify that the Temporary File Path field has the expected value.

Finally create a workflow with several connected steps and check that they execute on the Hub Engine.