Skip to content

Releases: stackhpc/ansible-slurm-appliance

v2.15

06 Mar 13:15
4595639

Choose a tag to compare

What's Changed

Security-related changes

  • Use upstream munge packages with fix for CVE-2026-25506 by @elelaysh in #912
    • NB: This CVE was previously fixed using an adhoc playbook to install StackHPC-built packages. This is no longer necessary as fixed upstream packages are included in the new images (see below).
  • Update jupyter to v1.1.1 to fix CVE-2025-7783 by @sjpb in #916
  • Update ondemand to v4.1.4 to fix GHSA-353f-x4gh-cqq8 by @sjpb in #918
  • Fix trivy warning for deepdiff package CVE-2025-58367 by @sjpb in #919

New features

  • Enable persistent journaling by @jovial in #763
  • Integrate opkssh for ssh using OIDC by @sjpb in #893

Other fixes and improvements

  • Fix permissions for sshd config when using compute-init by @jovial in #909
  • Fix secrets_openhpc_mungekey_default to be a plain value by @elelaysh in #907
  • Make filters to_ood_regex, prometheus_node_exporter_targets stable by @elelaysh in #900
  • Default NFS mounts to requiring working networking by @sjpb in #894
  • Run pulp CLI on pulp_server instead of localhost by @sjpb in #910
  • Configure MaxStartups in sshd and set for control by @MoteHue in #901
  • Use new OpenTofu-based jumphost for Leafcloud CI by @sjpb in #886

Images

Two new images are available:

  • RockyLinux8: openhpc-RL8-260306-0934-45b9a589
  • RockyLinux9: openhpc-RL9-260306-0934-45b9a589

Full Changelog: v2.14...v2.15

v2.14

18 Feb 10:45
72ec1e7

Choose a tag to compare

Important

The new images include our build of munge 0.5.18 to fix CVE-2026-25506.
As a precaution, you should rotate MUNGE keys in your cluster. Please see instructions here

What's Changed

  • Configurable administrator user groups for ssh by @elelaysh in #899
  • Provide fix for CVE-2026-25506: MUNGE buffer overflow allowing key leakage by @elelaysh in #903
  • Update StackHPC images to fix CVE-2026-25506 (upgrade to munge 0.5.18) by @elelaysh in #906
  • Upgrade pillow to 12.1.1 by @elelaysh in #908

Images

Two new images are available:

  • RL8: openhpc-RL8-260217-0837-79ae25d1
  • RL9: openhpc-RL9-260217-0837-79ae25d1

Full Changelog: v2.13...v2.14

v2.13

13 Feb 11:03
5ddd87f

Choose a tag to compare

What's Changed

  • Define cluster_subnets for caas by @sjpb in #904

Full Changelog: v2.12...v2.13

Images

There are no new images at this release, see v2.12

v2.12

12 Feb 12:25
d81afe8

Choose a tag to compare

What's Changed

  • Fix CI rebuild check for cases where image hasn't changed since latest release by @sjpb in #895
  • Ensure tmp.mount mask status matches fstab by @sjpb in #896 (Fixes failure for RL8 instances to complete reboot)
  • Fix OpenSSL bug RLSA-2026:1473 by @claudia-lola in #898 (NB: includes OS package update)
  • Bump python cryptography for CVE-2026-26007 by @sjpb in #902

Full Changelog: v2.11...v2.12

Images

Two new images are available:

  • RL8: openhpc-RL8-260210-0941-ccf0e76b
  • RL9: openhpc-RL9-260210-0941-ccf0e76b

v2.11

29 Jan 17:40
b118724

Choose a tag to compare

What's Changed

Important

By default, the control node will now run a squid service as a proxy for EESSI clients. See #876 and review the defaults in docs/eessi.md#eessi-proxy-configuration to ensure that they are suitable for your situation and control node sizing.

Updates

  • Update DNF packages & upgrade to Rocky Linux 9.7 by @elelaysh in #858
  • Bump CUDA to 13.1.1 and NVIDIA driver to 590.48.01 by @priteau in #880
  • Enable EESSI proxy by default by @sjpb in #876

New features and improvements

  • Re-enable gpgchecks for dnf packages, Manila patchlevel bump & docs improvements by @bertiethorpe in #873
  • Use Ark repos for Open Ondemand installs (with in-repo GPG key) by @sjpb in #831
  • alertmanager: support tuning systemd sandboxing by @priteau in #877
  • Add DEX support to openondemand role by @sjpb in #862
  • Add mounts role with /tmp defaulted to tmpfs on login & compute nodes by @sjpb in #888
  • Install bash-completion, for slurm command completions by @elelaysh in #892

Fixes and CI changes

  • Fix permissions issues for slurm-controlled rebuilds by @sjpb in #889
  • Bump python version for current release in CI by @sjpb in #885
  • Enable yamllint + key ordering + refresh super-linter by @elelaysh in #874
  • Provide useful error from topology if no instances found by @sjpb in #884

Images

Two new images are available:

  • RockyLinux 8: openhpc-RL8-260127-1007-d7ed9234
  • RockyLinux 9: openhpc-RL9-260127-1007-d7ed9234

Full Changelog: v2.10...v2.11

v2.10

31 Dec 14:03
b7109bd

Choose a tag to compare

What's Changed

Updates

Important

Python 3.12 is now required on the deploy host. The venv/ will be upgraded automatically on running dev/setup-env.sh (with no venv activated).

New features and improvements

Important

Running the site.yml playbook in production environments will prompt for confirmation by default - see PR883 below for details

  • Write summary from hpctests pingpong to repo by @sjpb in #870
  • Open OnDemand: Add dropdown to select select partition by @elelaysh in #869
  • Support templating and ansible_ssh_common_args in ansible-ssh by @sjpb in #872
  • Protect production environments when running Ansible by @sjpb in #883

Fixes & CI changes

  • Fix appliances_extra_packages_default missing stackhpc fatimage by @elelaysh in #856
  • Fix empty nodelist in ondemand apps by @sjpb in #861
  • Attempt to fix permissions for trivy scan on main by @JohnGarbutt in #867
  • Fix sssd & sshd for slurm-controlled rebuild by @sjpb in #866
  • .caas: fix missing path in docs for ANSIBLE_INVENTORY by @elelaysh in #865
  • CI: Cleanup OpenStack resources from previous attempts by @sjpb in #875
  • Fix TOFU ssh key prompt in OpenOnDemand web shell for IPA hosts by @sjpb in #864
  • alertmanager: Fix Slack integration by @priteau in #878
  • Update tf s3 backend instructions for better ec2 credential behaviour by @sjpb in #871
  • Fix hostkeys in IPA not matching host when persisting keys by @sjpb in #863
  • Fix Markdown formatting issues by @priteau in #881

Full Changelog: v2.9...v2.10

Images

Two new images are available:

  • RL8: openhpc-RL8-251213-1133-31273766
  • RL9: openhpc-RL9-251213-1133-31273766

v2.9

05 Dec 10:23
406d94b

Choose a tag to compare

What's Changed

Images

Two new images are available:

  • RockyLinux 8: openhpc-RL8-251119-1833-cb477455
  • RockyLinux 9: openhpc-RL9-251119-1834-cb477455

New Contributors

Full Changelog: v2.8.2...v2.9

v2.8.2

21 Nov 09:35
820645d

Choose a tag to compare

What's Changed

This is another minor update on v2.8 to avoid unnecessary delete/recreate of compute/login/control nodes:

  • Reformat user_data for other nodes by @sjpb in #851

For images see release v2.8.

Full Changelog: v2.8.1...v2.8.2

v2.8.1

19 Nov 16:54
9a048f6

Choose a tag to compare

What's Changed

This is a minor update on v2.8 to avoid unnecessary delete/recreate of compute nodes:

  • Avoid compute node replacements due to optional user_data by @MoteHue in #847
  • Use 'nodegroups' param in controlled rebuild docs by @MoteHue in #849

For images see release v2.8.

Full Changelog: v2.8...v2.8.1

v2.8

19 Nov 16:36
adb62d0

Choose a tag to compare

What's Changed

Improved GPU configuration/support

  • Support automatic GRES configuration for NVIDIA GPUs by @sjpb in #820
  • Add option to install nvidia-fabricmanager by @claudia-lola in #836
  • Add support for GRES to ondemand apps by @sjpb in #837
  • Adds bandwidth.yml playbook for NVIDIA nvbandwidth by @claudia-lola in #834
  • Make eessi configure gpu node automatically by @claudia-lola in #841
  • Bump CUDA to 13.0.2 and NVIDIA driver to 580.105.08 by @priteau in #823

Slurm configuration

New/improved features

  • Add support for InfiniBand interfaces to NHC by @sjpb in #821
  • Add tool to set image properties by @sjpb in #829

Docs and other

Full Changelog: v2.7...v2.8

Images

Two new images are available:

  • RL8: openhpc-RL8-251119-1202-332ac921
  • RL9: openhpc-RL9-251119-1202-332ac921