aws AWS What's New ·

AWS Parallel Computing Service enables in-place Slurm version upgrades

infraawsgaengineer
feature

AWS Parallel Computing Service (PCS) now supports managed in-place Slurm major version upgrades for existing clusters. This allows users to upgrade up to three major versions without disrupting running jobs, simplifying HPC workload management on AWS. The service automatically handles upgrades for controller, accounting database, and REST API components, preserving job continuity and accounting data. This feature is available in all AWS Regions where PCS is offered.

  • In-place Slurm major version upgrades supported for AWS Parallel Computing Service
  • Automatic upgrade of managed Slurm components
  • User convenience for compute node updates and documentation availability
Features (1)
  • In-place Slurm major version upgrades supported for AWS Parallel Computing Service

    AWS Parallel Computing Service (PCS) now supports managed in-place Slurm version upgrades for existing clusters. This feature allows moving clusters up to three Slurm major versions ahead with no disruption to running jobs.

Enhancements (1)
  • Automatic upgrade of managed Slurm components

    PCS handles the upgrade of all managed Slurm components, including the controller, accounting database, and REST API. Running jobs continue uninterrupted during the upgrade, and accounting data is preserved.

Notes (1)
  • User convenience for compute node updates and documentation availability

    Users can update their compute nodes to the new Slurm version at their convenience after the main cluster upgrade. More information on steps and considerations is available in the PCS User Guide.

Read the original announcement →

https://aws.amazon.com/about-aws/whats-new/2026/06/aws-parallel-computing-service-upgrade/