aws-eda-slurm-cluster v2.11.0

What's Changed

• Add support for ParallelCluster 3.12.0
• Change the RealMemory of compute nodes to match total instance type memory to prevent users from accidentally requesting twice as much memory as they really want because they don't specify 95% of actual memory.
• Document how to configure Slurm accounting.
• Fix XIO bugs, enhance configuration support
• Update RES templates for latest RES version. Update keycloak instance type to c7a.medium instead of t3.micro for stability.

New Features

• Feature #272: Improve documentation for ClusterConfig section
• Feature #275: Update RES templates for latest version
• Feature #277: Enable spot for only certain InstanceTypes
• Feature #282: document command11SubmitterDeconfigure command in instructions.
• Feature #283: Change RealMemory of compute nodes to match total instance type memory
• Feature #295: Add support for ParallelCluster 3.12.0

Bug Fixes

• Bug #253: ParallelCluster incorrectly requiring FSxZ egress rules
• Bug #280: Unable to create 3.10.1 cluster
• Bug #288: External login node can't acces slurmdbd
• Bug #293: Use of uninitialized variable

Full Changelog: v2.10.0...v2.11.0

aws-eda-slurm-cluster v2.10.0

What's Changed

• Add support for ParallelCluster 3.11.1.
• Add support for Exostellar Infrastructure Optimizer
• Update default instance types to use latest instance types. Use more instance types if only using on-demand or spot.
• Update lambdas from Python 3.9 to 3.12.

New Features

• Feature #226: Add Exostellar support
• Feature #268: Add support for ParallelCluster 3.11.1

Bug Fixes

• Bug #266: Document which python program needs to be updated if I need to create the user/group json file a different way
• Bug #267: slurmctld log shows "error: Node XXXXXXXXXX appears to have a different slurm.conf than the slurmctld."
• Fix bug in XIO configuration
• Require instances to have at least 4GB of memory.

Full Changelog: v2.9.0...v2.10.0

aws-eda-slurm-cluster v2.9.0

What's Changed

• Change the names of the compute resources to include the instance type and also the number of cores and amount of memory. This makes it easier for users to select compute nodes to use for their jobs.

New Features

• Add UseOnDemand configuration option that is similar to UseSpot so you can configure a cluster without on-demand instances if you want.

Bug Fixes

• Bug #261: Restore memory based partitions
• Bug #262: Default excludes incorrect if not using default includes
• Bug #264: Create partitions with number of cores and amount of memory in name

Full Changelog: v2.8.0...v2.9.0

aws-eda-slurm-cluster v2.8.0

New Features

• #258: Add support for ParallelCluster 3.11.0

v2.7.1

What's Changed

• Clean up security groups and permissions for extra mounts by @cartalla in #246
• Update deployment-prerequisites.md by @cartalla in #247

Full Changelog: v2.7.0...v2.7.1

aws-eda-slurm-cluster v2.7.0

What's Changed

• Add ParallelCluster 3.10.0, 3.10.1 support by @cartalla in #244

New Features

• Feature #242: Add support for ParallelCluster 3.10.0
• Feature #243: Add support for ParallelCluster 3.10.1

Bug Fixes

• Bug #221: Running install.sh with -cdk-cmd update in rapid succession can damage the cluster

Full Changelog: v2.6.0...v2.7.0

aws-eda-slurm-cluster v2.6.0

What's Changed

• Update deployment docs by @cartalla in #234
• Do not auto-prune instance types if there are too many by @cartalla in #235
• Support ParallelCluster 3.9.2 and 3.9.3. Fix ansible playbooks. by @cartalla in #241

New Features

• Feature #236: Add support for ParallelCluster 3.9.2
• Feature #240: Add support for ParallelCluster 3.9.3

Bug Fixes

• Bug #220: reducing number of compute resources to aggressively.
• Bug #222: Documentation corrections required on deploy-parallel-cluster documentation
• Bug #238: HeadNode fails to configure due to ansible change. on_head_node_configured.sh fails as ansible has deprecated ansible.builtin.include
• Bug #239: Documentation update: location of licenses is incorrect on doc page

Full Changelog: v2.5.0...v2.6.0

aws-eda-slurm-cluster v2.5.0

What's Changed

• Add support for ParallelCluster versions 3.9.0 and 3.9.1 by @cartalla in #232

New Features

• Feature #229: Add support for ParallelCluster version 3.9.0 and 3.9.1

Bug Fixes

• Bug #204: Can only configure 3 clusters on a submitter host
• Bug #230: Python 3.8 Lambda deprecated on 10/12/2024
  Update lambdas to use new version of python
• Bug #231: Cluster fails to deploy because create_slurm_accounts.py fails

Full Changelog: v2.4.0...v2.5.0

aws-eda-slurm-cluster v2.4.0

What's Changed

Add the following config options:

• slurm/ParallelClusterConfig/ClusterConfig
• slurm/SlurmCtl/AdditionalSecurityGroups
• slurm/SlurmCtl/AdditionalIamPolicies
• slurm/SlurmCtl/Imds/Secured
• slurm/InstanceConfig/AdditionalSecurityGroups
• slurm/InstanceConfig/AdditionalIamPolicies

Added documentation for all config parameters.

Changed the StackName default from slurm-top to slurm-config.

Fix the slurm/ParallelClusterConfig/Dcv/Enabled option.
Change the option name from Enable to Enabled to match ParallelCluster.

Fix the setting of ParallelCluster HeadNode/Dcv/AllowedIps config
Was setting from non-existent slurm/ParallelClusterConfig/AllowedIps instead of slurm/ParallelClusterConfig/HeadNode/Dcv/AllowedIps.

Delete the following config option because it uses legacy cluster.

• slurm/EdaSlurmClusterStackName

New Features

• Feature #225: Add custom IAM policies and security groups for head and compute
  Add config options for extra security groups and iam policies for hea… by @cartalla in #228

Full Changelog: v2.3.4...v2.4.0

aws-eda-slurm-cluster v2.3.4

What's Changed

New Features

• Feature #219: Update documentation for custom AMIs

Bug Fixes

• Bug #212: PyYAML 5.4.1 in source/requirements.txt does not install due to release of cython3.0
  Relax PyYAML version requirement by @cartalla in #215
• Bug #216: Delete local build files that can contain tokens or stale values
  Remove creation of local AMI build-files by @cartalla in #217
• Bug #223: module load sets environment variables that override values in the sbatch submission script
  Remove sbatch and srun defaults from modulefile by @cartalla in #224

Full Changelog: v2.3.3...v2.3.4