AWS Parallel Computing Service adds node lifecycle actions
AWS Parallel Computing Service (PCS) is now generally available with node lifecycle actions, allowing custom scripts to run automatically at defined points in a compute node's lifecycle. This enables automated node preparation, such as mounting storage or installing software, for HPC workloads managed by Slurm. The feature is available in all AWS PCS regions and provides logging for visibility.
- →Automate compute node setup with lifecycle actions
- →AWS PCS simplifies HPC workload management
Features (1) ›
- Automate compute node setup with lifecycle actions
AWS Parallel Computing Service (PCS) now supports node lifecycle actions, allowing custom scripts to run automatically during node setup, configuration, or reboot. Scripts can be defined with parameters, lifecycle stages, and error handling, and their execution is logged for visibility. This feature simplifies preparing compute nodes for HPC workloads.
Notes (1) ›
- AWS PCS simplifies HPC workload management
AWS Parallel Computing Service is a managed service designed to simplify running and scaling high-performance computing (HPC) workloads on AWS using Slurm. It offers integrated compute, storage, networking, and visualization tools, with automated cluster management and observability.
https://aws.amazon.com/about-aws/whats-new/2026/07/aws-pcs-node-lifecycle-actions/
Related releases
- AWS Lambda adds direct read configuration for Amazon S3 Files AWS What's New ·
- SageMaker HyperPod adds model caching for faster inference autoscaling AWS What's New ·
- Amazon Bedrock Knowledge Base adds TwelveLabs Marengo 3.0 for multimodal embeddings AWS What's New ·
- AWS Lambda recursive loop detection now available in Europe Sovereign Cloud AWS What's New ·
- AWS Storage Gateway adds FIPS-compliant private connectivity for S3 File Gateway AWS What's New ·
- Amazon API Gateway now supports 1 MB execution logs with configurable delivery destinations AWS What's New ·