Cluster Toolkit is an open-source solution that simplifies deploying
high performance computing (HPC), artificial intelligence (AI), and machine
learning (ML) workloads on Google Cloud. With customizable blueprints, you can
provision infrastructure that aligns with Google Cloud best practices, which
helps to ensure an efficient and consistent setup for your compute environments.
The gcluster command-line tool helps you deploy, review, and monitor these
configured infrastructure blueprints. Although you can update some active cluster
configurations, making fundamental architectural changes might require
redeploying the cluster.
Before you begin
- If you are using a Linux or macOS workstation to deploy your cluster, then follow install dependencies.
- From either your workstation or Cloud Shell, configure your environment.
- Verify that you have created a cluster blueprint or selected one from the Cluster Toolkit blueprint catalog.
Deploy the cluster
To deploy your cluster using a blueprint, complete the following steps:
Switch to the main Cluster Toolkit working directory:
cd cluster-toolkitDeploy the cluster using the
gcluster deploycommand:To deploy using command-line variables:
./gcluster deploy --vars project_id=PROJECT_ID PATH_TO_BLUEPRINT_FILETo deploy with a deployment file:
./gcluster deploy -d PATH_TO_DEPLOYMENT_FILE PATH_TO_BLUEPRINT_FILE
Replace the following:
PROJECT_ID: The unique identifier for your target Google Cloud project.PATH_TO_BLUEPRINT_FILE: The relative or absolute path to your cluster blueprint file.PATH_TO_DEPLOYMENT_FILE: The path to your cluster deployment configuration file.
The
gclustercommand reports proposed changes for your cluster. To continue the deployment, accept the proposed changes by pressing a, and then pressing Enter. Optionally, you can review the proposed changes by pressing d, and then pressing Enter.Summary of proposed changes: Plan: 37 to add, 0 to change, 0 to destroy. (D)isplay full proposed changes, (A)pply proposed changes, (S)top and exit, (C)ontinue without applying Please select an option [d,a,s,c]: aAfter accepting the changes,
gclusterrunsterraform applyautomatically. This operation takes approximately five minutes while it displays progress. If the run is successful, the output is similar to the following:Apply complete! Resources: 37 added, 0 changed, 0 destroyed.
Get command-line help
To view all available flags and commands, run the --help flag:
./gcluster --help
./gcluster deploy --help
What's next
- Review Best practices for HPC workloads.
- Learn how to manage active deployments in Reconfigure a running cluster.