Run jobs using the SageMaker HyperPod CLI

To run jobs, make sure that you installed Kubeflow Training Operator in the EKS clusters. For more information, see Install packages on the Amazon EKS cluster using Helm.

Run the hyperpod get-cluster command to get the list of available HyperPod clusters.

hyperpod get-clusters

Run the hyperpod connect-cluster to configure the SageMaker HyperPod CLI with the EKS cluster orchestrating the HyperPod cluster.

hyperpod connect-cluster --cluster-name <hyperpod-cluster-name>

Use the hyperpod start-job command to run a job. The following command shows the command with required options.

hyperpod start-job \
    --job-name <job-name>
    --image <docker-image-uri>
    --entry-script <entrypoint-script>
    --instance-type <ml.instance.type>
    --node-count <integer>

The hyperpod start-job command also comes with various options such as job auto-resume and job scheduling.

Enabling job auto-resume

The hyperpod start-job command also has the following options to specify job auto-resume. For enabling job auto-resume to work with the SageMaker HyperPod node resiliency features, you must set the value for the restart-policy option to OnFailure. The job must be running under the kubeflow namespace or a namespace prefixed with hyperpod.

• [--auto-resume <bool>] #Optional, enable job auto resume after fails, default is false

• [--max-retry <int>] #Optional, if auto-resume is true, max-retry default value is 1 if not specified

• [--restart-policy <enum>] #Optional, PyTorchJob restart policy. Available values are Always, OnFailure, Never or ExitCode. The default value is OnFailure.

hyperpod start-job \
    ... // required options \
    --auto-resume true \
    --max-retry 3 \
    --restart-policy OnFailure

Running jobs with scheduling options

The hyperpod start-job command has the following options to set up the job with queuing mechanisms.

Note

You need Kueue installed in the EKS cluster. If you haven't installed, follow the instructions at Install packages on the Amazon EKS cluster using Helm.

• [--scheduler-type <enum>] #Optional, Specify the scheduler type. The default is Kueue.

• [--queue-name <string>] #Optional, Specify the name of the Local Queue or Cluster Queue you want to submit with the job. The queue should be created by cluster admins.

• [--priority <string>] #Optional, Specify the name of the Workload Priority Class, which should be created by cluster admins.

hyperpod start-job \
    ... // required options
    --scheduler-type Kueue \
    --queue-name high-priority-queue \
    --priority high

Running jobs from a configuration file

As an alternative, you can create a job configuration file containing all the parameters required by the job and then pass this config file to the hyperpod start-job command using the --config-file option. In this case:

1. Create your job configuration file with the required parameters. Refer to the job configuration file in the SageMaker HyperPod CLI GitHub repository for a baseline configuration file.

2. Start the job using the configuration file as follows.

   hyperpod start-job --config-file /path/to/test_job.yaml

Tip

For a complete list of parameters of the hyperpod start-job command, see the Submitting a Job section in the README.md of the SageMaker HyperPod CLI GitHub repository.