AWS pcs documentation change
Summary
Added scheduler version to modifiable cluster configurations and included new FAQ section with instructions for updating standalone login nodes during scheduler version updates. Removed Slurm version from modifiable items.
Security assessment
Operational update about cluster management procedures. No security vulnerabilities are mentioned, and changes focus on update mechanics rather than security features.
Diff
diff --git a/pcs/latest/userguide/working-with_clusters_update_faq.md b/pcs/latest/userguide/working-with_clusters_update_faq.md index 17029720c..ad905afe5 100644 --- a//pcs/latest/userguide/working-with_clusters_update_faq.md +++ b//pcs/latest/userguide/working-with_clusters_update_faq.md @@ -14 +14 @@ Get answers to common questions about updating cluster configurations in AWS PCS -You can modify accounting configuration (enable/disable managed accounting), scale-down behavior (scaleDownIdleTime parameter), and any of the supported Slurm custom settings that apply at the cluster level. You cannot modify security groups, VPC subnets, cluster size, Slurm version, or cluster name. +You can modify accounting configuration (enable/disable accounting), scale-down behavior (scaleDownIdleTime parameter), the scheduler version, and any of the supported Slurm custom settings that apply at the cluster level. You cannot modify security groups, VPC subnets, cluster size, or cluster name. @@ -35,0 +36,5 @@ Yes, running jobs continue to execute on compute nodes even when the cluster con +**How do I update standalone login nodes during a scheduler version update?** + + +Standalone login nodes are EC2 instances that are not managed by AWS PCS compute node groups. During a scheduler version update, AWS PCS does not automatically update these instances. You must update them manually by replacing them with new instances that use an AMI containing the target Slurm version and the latest AWS PCS agent. For Option 1 (rolling update), update login nodes after the controller update is complete and the fleet is fully on the new version. For Option 2 (full-fleet recycle), you can update login nodes at any point during the process. For more information, see [Update the scheduler version of an AWS PCS cluster](./working-with_clusters_version_update_procedure.html). + @@ -47 +52 @@ To use the Amazon Web Services Documentation, Javascript must be enabled. Please -Update cluster +Troubleshooting version updates