SageMaker HyperPod simplifies Slurm cluster setup with AMI-based configuration
Amazon SageMaker HyperPod now supports AMI-based node lifecycle configuration for Slurm clusters using continuous provisioning. This eliminates the need for manual script uploads, speeding up node availability for AI/ML workloads. The feature is available in all SageMaker HyperPod regions and targets engineers and architects managing ML infrastructure.
- →Enable AMI-based node lifecycle configuration for Slurm clusters with continuous provisioning
- →Streamlined node provisioning with pre-configured software and settings
- →Flexible configuration options via API or SageMaker console
- →Custom lifecycle configuration scripts remain fully supported
Features (1) ›
- Enable AMI-based node lifecycle configuration for Slurm clusters with continuous provisioning
SageMaker HyperPod now supports AMI-based configuration for Slurm clusters that use continuous provisioning. This simplifies setup by removing the need to download, configure, or upload lifecycle configuration scripts to Amazon S3, allowing nodes to become available for jobs sooner.
Enhancements (1) ›
- Streamlined node provisioning with pre-configured software and settings
Clusters using continuous provisioning can now be created with pre-configured AMIs that include essential software like Docker, Enroot, and Pyxis, along with configurations for Slurm accounting and SSH key generation. This ensures nodes are provisioned in a production-ready state for AI/ML training workloads.
Notes (2) ›
- Flexible configuration options via API or SageMaker console
AMI-based configuration can be enabled by omitting the LifeCycleConfig block via the API or selecting 'None' for Lifecycle scripts in the SageMaker AI console. For additional customization, extension scripts can be provided via S3 URI.
- Custom lifecycle configuration scripts remain fully supported
Users requiring full control over node provisioning can continue to use custom lifecycle configuration scripts. The new AMI-based option complements existing functionality for enhanced flexibility.
https://aws.amazon.com/about-aws/whats-new/2025/06/ami-configuration-continuous-slurm/