Merge branch 'main' into feat/eess-stratum-1

Author: sjpb

.github/workflows/fatimage.yml

@@ -31,7 +31,7 @@ jobs:
             source_image_name: Rocky-8-GenericCloud-Base-8.10-20240528.0.x86_64.raw
             inventory_groups: control,compute,login,update
           - image_name: openhpc-RL9
-            source_image_name: Rocky-9-GenericCloud-Base-9.5-20241118.0.x86_64.raw
+            source_image_name: Rocky-9-GenericCloud-Base-9.6-20250531.0.x86_64.qcow2
             inventory_groups: control,compute,login,update
     env:
       ANSIBLE_FORCE_COLOR: True

ansible/fatimage.yml

@@ -190,6 +190,11 @@
         tasks_from: install.yml
       when: "'alertmanager' in group_names"
 
+    - name: Download HPL source
+      include_role:
+        name: hpctests
+        tasks_from: source-hpl.yml
+
 - hosts: prometheus
   become: yes
   gather_facts: yes

ansible/roles/hpctests/tasks/build-hpl.yml

@@ -5,18 +5,21 @@
     path: "{{ hpctests_rootdir }}/hpl"
     state: directory
 
-- name: Download HPL sources
+- name: Unarchive HPL sources from /opt/hpl
   unarchive:
-    src: "http://www.netlib.org/benchmark/hpl/hpl-{{ hpctests_hpl_version }}.tar.gz"
-    remote_src: yes
+    src: "/opt/hpl/hpl-{{ hpctests_hpl_version }}.tar.gz"
     dest: "{{ hpctests_rootdir }}/hpl"
+    remote_src: yes
+    owner: "{{ hpctests_user }}"
+    group: "{{ hpctests_group }}"
+    mode: '0755'
     keep_newer: yes
 
-- name: Copy BLAS make file
-  command:
-    cmd: "cp setup/Make.Linux_PII_CBLAS Make.{{ hpctests_hpl_arch }}"
-    chdir: "{{ hpctests_hpl_srcdir }}"
-    creates: "{{ hpctests_hpl_srcdir }}/Make.{{ hpctests_hpl_arch }}"
+- name: Copy BLAS makefile
+  copy:
+    src: "{{ hpctests_hpl_srcdir }}/setup/Make.Linux_PII_CBLAS"
+    dest: "{{ hpctests_hpl_srcdir }}/Make.{{ hpctests_hpl_arch }}"
+    remote_src: yes
 
 - name: Modify make file
   replace:

ansible/roles/hpctests/tasks/source-hpl.yml

@@ -0,0 +1,17 @@
+---
+
+- name: Make directory
+  file:
+    path: "/opt/hpl"
+    state: directory
+    owner: root
+    group: root
+    mode: '0755'
+
+- name: Download HPL tarball
+  get_url:
+    url: "http://www.netlib.org/benchmark/hpl/hpl-{{ hpctests_hpl_version }}.tar.gz"
+    dest: "/opt/hpl/hpl-{{ hpctests_hpl_version }}.tar.gz"
+    owner: root
+    group: root
+    mode: '0644'

ansible/roles/lustre/README.md

@@ -19,10 +19,8 @@ The following variables control configuration of Lustre clients.
 - `lustre_mount_options`. Optional default mount options. Default values are systemd defaults from [Lustre client docs](http://wiki.lustre.org/Mounting_a_Lustre_File_System_on_Client_Nodes).
 
 The following variables control the package build and and install:
-- `lustre_version`: Optional str. Version of lustre to build, default `2.15.6/lu-18085`
-  which is the first version with EL9.5 support, plus a fix for https://jira.whamcloud.com/browse/LU-18085.
-- `lustre_repo`: Optional str. URL for Lustre repo. Default is a StackHPC repo
-  incorporating the above fix.
+- `lustre_version`: Optional str. Version of lustre to build, default `2.15.7`
+- `lustre_repo`: Optional str. URL for Lustre repo. Default is `git://git.whamcloud.com/fs/lustre-release`.git.
 - `lustre_build_packages`: Optional list. Prerequisite packages required to build Lustre. See `defaults/main.yml`.
 - `lustre_build_dir`: Optional str. Path to build lustre at, default `/tmp/lustre-release`.
 - `lustre_configure_opts`: Optional list. Options to `./configure` command. Default builds client rpms supporting Mellanox OFED, without support for GSS keys.

ansible/roles/lustre/defaults/main.yml

@@ -1,13 +1,12 @@
-lustre_repo: https://github.com/stackhpc/lustre-release.git
-lustre_version: '2.15.6/lu-18085' # Fixes https://jira.whamcloud.com/browse/LU-18085
+lustre_version: '2.15.7'
 lustre_lnet_label: tcp
 #lustre_mgs_nid:
 lustre_mounts: []
 lustre_mount_state: mounted
 lustre_mount_options: 'defaults,_netdev,noauto,x-systemd.automount,x-systemd.requires=lnet.service,nosuid,nodev'
 
 # below variables are for build and should not generally require changes
-lustre_git_repo: "git://git.whamcloud.com/fs/lustre-release.git"
+lustre_repo: "git://git.whamcloud.com/fs/lustre-release.git"
 lustre_build_packages:
   - "kernel-devel-{{ ansible_kernel }}"
   - git

ansible/validate.yml

@@ -64,14 +64,6 @@
     - import_role:
         name: stackhpc.openhpc
         tasks_from: validate.yml
-    - assert:
-        that: "'enable_configless' in openhpc_config.SlurmctldParameters | default([])"
-        fail_msg: |
-          'enable_configless' not found in openhpc_config.SlurmctldParameters - is variable openhpc_config overridden?
-          Additional slurm.conf parameters should be provided using variable openhpc_config_extra.
-        success_msg: Checked Slurm will be configured for configless operation
-      delegate_to: localhost
-      run_once: true
 
 - name: Validate filebeat configuration
   hosts: filebeat

docs/experimental/isolated-clusters.md

@@ -12,14 +12,7 @@ created by cookiecutter for a new environment.
 The full list of features and whether they are functional on such an "isolated"
 network is shown in the table below. Note that:
 
-1. The `hpl` test from the `ansible/adhoc/hpctests.yml` playbook is not
-   functional and must be skipped using:
-
-    ```shell
-    ansible-playbook ansible/adhoc/hpctests.yml --skip-tags hpl-solo
-    ```
-
-2. Using [EESSI](https://www.eessi.io/docs/) necessarily requires outbound
+-  Using [EESSI](https://www.eessi.io/docs/) necessarily requires outbound
    network access for the CernVM File System. If security groups are not
    sufficent to restrict this:
     a. If outbound http is available, an authenticated proxy could be used,
@@ -54,7 +47,7 @@ See above for definition of "Default" features. In the "Isolated?" column:
 | freeipa_client        | - | Y - image build required |
 | gateway               | n/a | n/a - build only | 
 | grafana               | Y | Y | 
-| hpctests              | Y | Y  - except hpl-solo, see above | 
+| hpctests              | Y | Y | 
 | k3s_agent             | - | ? | 
 | k3s_server            | - | ? | 
 | k9s                   | - | ? | 

environments/.stackhpc/tofu/cluster_image.auto.tfvars.json

@@ -1,6 +1,6 @@
 {
     "cluster_image": {
-        "RL8": "openhpc-RL8-250708-1502-1494192e",
-        "RL9": "openhpc-RL9-250708-1547-1494192e"
+        "RL8": "openhpc-RL8-250805-1410-35724c15",
+        "RL9": "openhpc-RL9-250805-1409-35724c15"
     }
 }

environments/common/inventory/group_vars/all/openhpc.yml

@@ -43,7 +43,7 @@ openhpc_packages_default:
   - podman-compose
 openhpc_packages_extra: []
 openhpc_packages: "{{ (openhpc_packages_default + openhpc_packages_extra) | select | list }}"
-openhpc_munge_key: "{{ vault_openhpc_mungekey | b64decode }}"
+openhpc_munge_key_b64: "{{ vault_openhpc_mungekey }}"
 openhpc_login_only_nodes: login
 openhpc_state_save_location: "{{ appliances_state_dir + '/slurmctld' if appliances_state_dir is defined else '/var/spool' }}"
 
@@ -53,8 +53,6 @@ openhpc_config_extra: {}
 
 # default additional slurm.conf parameters for the appliance:
 openhpc_config_default:
-  SlurmctldParameters:
-    - enable_configless
   TaskPlugin: task/cgroup,task/affinity
   ReturnToService: 2 # workaround for templating bug TODO: Remove once on stackhpc.openhpc v1.2.0
   TopologyPlugin: "topology/{{ 'tree' if (topology_nodes | length) > 0 else 'flat' }}"