IBM Cloud Docs
Getting started with IBM Spectrum Symphony

Getting started with IBM Spectrum Symphony

IBM® Spectrum Symphony enables customers to deploy HPC clusters that use IBM Spectrum Symphony as a scheduling software. The deployment is performed by using Terraform and IBM Cloud Schematics as automation frameworks. The following steps outline the high-level flow of events that are performed:

  1. Create a workspace with the Terraform code from IBM Cloud Schematics. This step defines the set of configuration properties that are used to perform the automation.
  2. Generate a plan to confirm whether the configuration properties are valid, so that when you run the Terraform code, all of the resources are provisioned correctly. If the validation fails, fix the configuration properties and try again.
  3. Apply a plan triggers the actual deployment of the IBM Cloud resources to have an HPC cluster up and running by the time the deployment completes. If the deployment fails, identify the reason for failure, fix the problem, and try again. If a change is needed to the configuration properties, it might be better to generate a plan again.

If you decide to deploy your IBM Spectrum Symphony cluster through the IBM Cloud catalog, when you click Install, the Generate Plan action is skipped, and the steps go from Create Workspace to Apply Plan directly. You need to enter values in the catalog that work for your permissions and IBM Cloud account. If the deployment fails, the Schematics UI can be used to fix the errors, and you can retry the Apply Plan step.

Before you can deploy your Spectrum Symphony cluster, you need to create or gather some information. To get started, complete the following steps.

Generate API key

Generate an API key for your IBM Cloud account where the Spectrum Symphony cluster will be deployed. For more information, see Managing user API keys.

Create SSH key

Create an SHH key in your IBM Cloud account. This is your SSH key that you will use to access the Symphony cluster. For more information, see Managing SSH keys.

Create custom image(s)

The offering provides a default set of images that you can use for the nodes within your HPC cluster. However, if you would prefer to use your own custom images, documentation and scripts are provided to help you create them.

Worker Image: The default image that is specified in image_name acts as the worker image. This image is used to create the Symphony cluster nodes: management, management-candidates, and workers.

If you prefer to create the worker custom image with your own settings and configurations, follow the instructions that are provided here.

Storage Image: If you intend to use Storage Scale as shared storage for the cluster, the default image specified in scale_storage_image_name acts as the storage image. This image is used to create the Storage Scale storage cluster nodes. In addition to the base operating system, the image also includes the required Storage Scale software packages.

If you prefer to create the storage custom image with your own settings and configurations, follow the instructions that are provided here.

Gather Spectrum software entitlement information

The offering uses the Bring Your Own License (BYOL) model for Spectrum software when you deploy an HPC cluster on IBM Cloud. This applies to IBM Spectrum Symphony in all cases and to IBM Storage Scale if you decide to use that for the HPC cluster storage instead of just NFS. Work with your business owners or license management team to make sure that your organization has procured enough licenses to deploy the HPC cluster using the Spectrum software. Failure to comply with licenses for production use of software is a violation of the IBM International Program License Agreement.

Next steps

After you've gathered the necessary input values to define your cluster configuration, you are ready to deploy your IBM Spectrum Symphony cluster. The Spectrum Symphony cluster can be deployed on IBM Cloud by using the IBM Cloud catalog, Schematics CLI, or the Schematics APIs. If you want to deploy your cluster by using the CLI or API, review the prerequisites for your interface of choice:

After you've created and gathered your information and reviewed any additional prerequisites for your interface of choice, you are ready to begin Creating a workspace.