v1.146

Key Changes - all clusters

• Fix '400 HTTP Bad Request' when adding IPA hosts to in-cluster dev FreeIPA server by @sjpb in #389
• Define compute groups in OpenTofu by @sjpb in #390
• Restrict node resources by @sjpb in #388
• Bump openhpc role to fix legacy selecttype by @sjpb in #392

Key Changes - Azimuth-deployed clusters

• Add storage VLAN interface on all slurm nodes by @scrungus in #391

Deployment Notes

Roles/collections from Ansible Galaxy have changed. To update these run:

dev/setup-env.sh

Image Details

No new images are available.

New Contributors

• @scrungus made their first contribution in #391

Full Changelog: v1.145...v1.146

v1.145

What's Changed

• Use RockyLinux 9 for CI by default by @sjpb in #385
• Fix ControlPath for cookiecutter ansible.cfg by @sjpb in #386
• Add support for OFED by @sjpb in #254

Full Changelog: v1.144...v1.145

Deployment Notes

No changes to galaxy roles/collections.

Image Details

Two new images are available:

• RL8 without OFED: openhpc-RL8-240423-1002-4b09ba85 - requires a 12GB root disk
• RL9 with OFED 24.01-0.3.3.1: openhpc-ofed-RL9-240423-1059-4b09ba85 - requires a 15GB root disk

v1.144

Key Changes - all clusters

• Default OS is now RockyLinux 9:
  • Add support for RockyLinux9 by @sjpb in #353
  • Use RL9 for caas environment by @sjpb in #380
• "Enable cockpit" message of the day removed by @sjpb in #381
• basic_users role now supports setting sudo roles via #383

Key Changes - Azimuth-deployed clusters

• azimuth user now has sudo on all nodes by @sjpb in #383
• Control nodes now require >=4GB memory to avoid memory issues (#378, #379)
• HPL test memory target lowered from 80% to 5% memory for faster deployment by @sjpb in #384

CI- and image-build related changes

• Remove special-case for firewalld in builder by @sjpb in #364
• Enable CI on leafcloud by @sjpb in #369
• Improve leafcloud CI support by @sjpb in #374
• Fix leafcloud fatimage build by @sjpb in #376
• Fix home directory for rocky during image cleanup by @sjpb in #382

Other changes

• Fix freeipa server referer errors by @sjpb in #372
• Fix filebeat service reload by @sjpb in #373
• Add hpctests_pre_cmd option by @sjpb in #375

Full Changelog: v1.143...v1.144

Deployment Notes

Roles/collections from Ansible Galaxy have changed. To update these run:

dev/setup-env.sh

Image Details

Two new images are available, both requiring a 12GB root disk:

• RL8: openhpc-RL8-240327-1050-4812f852
• RL9: openhpc-RL9-240327-1026-4812f852

v1.143

What's Changed

• Use OpenTofu instead of Terraform for CaaS by @sjpb in #362

Full Changelog: v1.142...v1.143

Deployment notes

No changes to galaxy roles/collections.

Image Details

No new image provided by this release.

v1.142

What's Changed

• Don't link manila for /home to manila for /project in caas by @sjpb in #363

Full Changelog: v1.141...v1.142

Deployment notes

No changes to galaxy roles/collections.

Image Details

No new image provided by this release.

v1.141

What's Changed

• Use project manila share user from Azimuth by @sjpb in #359
• Address CI comments by @sjpb in #347
• Fix failure when setting basic_users state: absent by @sjpb in #361

Full Changelog: v1.140.1...v1.141

Deployment notes

No changes to galaxy roles/collections.

Image Details

No new image provided by this release.

v1.140.1

What's Changed

• Add manila /home specific UI for Arcus by @sjpb in #357
• Require user SSH key to be available in ui-meta by @JohnGarbutt in #356
• Improve CaaS UI descriptions by @sjpb in #358

Full Changelog: v1.140...v1.140.1

These changes only affect the Azimuth CaaS deployments. There are no changes to galaxy roles/collections or images.

v1.140

What's Changed

• Changes to support explicit quota checks (in Azimuth/.caas environment) by @mkjpryor in #348
• Remove the cluster_previous_image workaround (in Azimuth/.caas environment) by @JohnGarbutt in #355
• Support manila fileshares (cephfs) by @sjpb in #344. This adds both general support, and also default configuration for the Azimuth / .caas environment. See linked PR for full details.

Full Changelog: v.139...v1.140

Deployment notes

The stackhpc.os-manila-mount has been added. Run

dev/setup-env.sh

to install this

Image details

New image openhpc-240116-1604-b3563a08 available, requires 12GB root disk

v1.139

What's Changed

• Pull containers by @sjpb in #351:
  - Container images pulled before service start and during fat image build.
  - Fixes an issue where podman commands failed after reboot.
  See PR for full details.
• Use most-recent image in skeleton terraform if multiple found by @sjpb in #350

Full Changelog: v1.138...v.139

Deployment notes

No changes to galaxy roles/collections.

Image Details

New fat image openhpc-240116-1156-aa8dba7d, requires 12GB root disk.

CI changes

• FIP now used for build on Arcus to avoid docker.io rate limits.

v1.138

What's Changed

• Don't ignore image changes in skeleton terraform lifecycle by @sjpb in #334
• Cope with stale NFS file handles by @sjpb in #332
• Update fatimage base to RL8.9 with robust volume mounts by @sjpb in #341
• Remove cve-2023-41914 role by @sjpb in #337
• Avoid prompting user to accept hostkey in OOD shell by @sjpb in #331
• Fix removal of packer bundled ansible plugin by @sjpb in #346

Full Changelog: v1.137...v1.138

Deployment notes

The stackhpc.nfs role has changed. To update this run:

dev/setup-env.sh

Image Details

New image openhpc-240102-1025-e533fd70, requiring 10GB root disk.