Merge branch 'main' into feat/tf-nodegroup-typedefs

Author: sjpb

ansible/roles/cluster_infra/templates/outputs.tf.j2

@@ -32,12 +32,12 @@ output "cluster_nodes" {
         }
       }
     ],
-    {% for partition in openhpc_slurm_partitions %}
+    {% for nodegroup in openhpc_nodegroups %}
     [
-      for compute in openstack_compute_instance_v2.{{ partition.name }}: {
+      for compute in openstack_compute_instance_v2.{{ nodegroup.name }}: {
         name          = compute.name
         ip            = compute.network[0].fixed_ip_v4
-        groups        = ["compute", "{{ cluster_name }}_compute", "{{ cluster_name }}_{{ partition.name }}"],
+        groups        = ["compute", "{{ cluster_name }}_compute", "{{ cluster_name }}_{{ nodegroup.name }}"],
         facts  = {
           openstack_project_id = data.openstack_identity_auth_scope_v3.scope.project_id
         }

ansible/roles/cluster_infra/templates/resources.tf.j2

@@ -282,11 +282,11 @@ resource "openstack_networking_port_v2" "control_storage" {
 ###
 # Workers
 ###
-{% for partition in openhpc_slurm_partitions %}
+{% for nodegroup in openhpc_nodegroups %}
 # Primary network
-resource "openstack_networking_port_v2" "{{ partition.name }}" {
-  count = {{ partition.count }}
-  name = "{{ cluster_name }}-compute-{{ partition.name }}-${count.index}"
+resource "openstack_networking_port_v2" "{{ nodegroup.name }}" {
+  count = {{ nodegroup.count }}
+  name = "{{ cluster_name }}-compute-{{ nodegroup.name }}-${count.index}"
   network_id = "${data.openstack_networking_network_v2.cluster_network.id}"
   admin_state_up = "true"
 
@@ -305,9 +305,9 @@ resource "openstack_networking_port_v2" "{{ partition.name }}" {
 
 # Storage network
 {% if cluster_storage_network is defined %}
-resource "openstack_networking_port_v2" "{{ partition.name }}_storage" {
-  count          = {{ partition.count }}
-  name           = "{{ cluster_name }}-compute-{{ partition.name }}-storage-${count.index}"
+resource "openstack_networking_port_v2" "{{ nodegroup.name }}_storage" {
+  count          = {{ nodegroup.count }}
+  name           = "{{ cluster_name }}-compute-{{ nodegroup.name }}-storage-${count.index}"
   network_id     = data.openstack_networking_network_v2.cluster_storage.id
   admin_state_up = "true"
 
@@ -499,25 +499,25 @@ resource "openstack_compute_instance_v2" "control" {
   }
 }
 
-{% for partition in openhpc_slurm_partitions %}
-resource "openstack_compute_instance_v2" "{{ partition.name }}" {
-  count = {{ partition.count }}
+{% for nodegroup in openhpc_nodegroups %}
+resource "openstack_compute_instance_v2" "{{ nodegroup.name }}" {
+  count = {{ nodegroup.count }}
 
-  name      = "{{ cluster_name }}-compute-{{ partition.name }}-${count.index}"
+  name      = "{{ cluster_name }}-compute-{{ nodegroup.name }}-${count.index}"
   image_id  = "{{ cluster_image }}"
-  {% if 'flavor_name' in partition %}
-  flavor_name = "{{ partition.flavor_name }}"
+  {% if 'flavor_name' in nodegroup %}
+  flavor_name = "{{ nodegroup.flavor_name }}"
   {% else %}
-  flavor_id = "{{ partition.flavor }}"
+  flavor_id = "{{ nodegroup.flavor }}"
   {% endif %}
 
   network {
-    port = openstack_networking_port_v2.{{ partition.name }}[count.index].id
+    port = openstack_networking_port_v2.{{ nodegroup.name }}[count.index].id
   }
 
   {% if cluster_storage_network is defined %}
   network {
-    port = openstack_networking_port_v2.{{ partition.name }}_storage[count.index].id
+    port = openstack_networking_port_v2.{{ nodegroup.name }}_storage[count.index].id
   }
   {% endif %}
 

ansible/roles/cuda/README.md

@@ -10,6 +10,6 @@ Requires OFED to be installed to provide required kernel-* packages.
 
 - `cuda_repo_url`: Optional. URL of `.repo` file. Default is upstream for appropriate OS/architecture.
 - `cuda_nvidia_driver_stream`: Optional. Version of `nvidia-driver` stream to enable. This controls whether the open or proprietary drivers are installed and the major version. Changing this once the drivers are installed does not change the version.
-- `cuda_packages`: Optional. Default: `['cuda', 'nvidia-gds', 'cmake', 'cuda-toolkit-12-8']`.
+- `cuda_packages`: Optional. Default: `['cuda', 'nvidia-gds', 'cmake', 'cuda-toolkit-12-9']`.
 - `cuda_package_version`: Optional. Default `latest` which will install the latest packages if not installed but won't upgrade already-installed packages. Use `'none'` to skip installing CUDA.
 - `cuda_persistenced_state`: Optional. State of systemd `nvidia-persistenced` service. Values as [ansible.builtin.systemd:state](https://docs.ansible.com/ansible/latest/collections/ansible/builtin/systemd_module.html#parameter-state). Default `started`.

ansible/roles/cuda/defaults/main.yml

@@ -1,12 +1,12 @@
 cuda_repo_url: "https://developer.download.nvidia.com/compute/cuda/repos/rhel{{ ansible_distribution_major_version }}/{{ ansible_architecture }}/cuda-rhel{{ ansible_distribution_major_version }}.repo"
-cuda_nvidia_driver_stream: '570-open'
-cuda_package_version: '12.8.1-1'
-cuda_version_short: '12.8'
+cuda_nvidia_driver_stream: '575-open'
+cuda_package_version: '12.9.0-1'
+cuda_version_short: '12.9'
 cuda_packages:
   - "cuda{{ ('-' + cuda_package_version) if cuda_package_version != 'latest' else '' }}"
   - nvidia-gds
   - cmake
-  - cuda-toolkit-12-8
+  - cuda-toolkit-12-9
 cuda_samples_release_url: "https://github.com/NVIDIA/cuda-samples/archive/refs/tags/v{{ cuda_version_short }}.tar.gz"
 cuda_samples_path: "/var/lib/{{ ansible_user }}/cuda_samples"
 cuda_samples_programs:

ansible/roles/cuda/tasks/samples.yml

@@ -25,36 +25,3 @@
     cmd: . /etc/profile.d/sh.local && cmake .. && make -j {{ ansible_processor_vcpus }}
     chdir: "{{ cuda_samples_path }}/cuda-samples-{{ cuda_version_short }}/build"
     creates: "{{ cuda_samples_path }}/cuda-samples-{{ cuda_version_short }}/build/Samples/1_Utilities/deviceQuery/deviceQuery"
-
-- name: Run CUDA deviceQuery
-  command:
-    cmd: "{{ cuda_samples_path }}/cuda-samples-{{ cuda_version_short }}/build/Samples/1_Utilities/deviceQuery/deviceQuery"
-  register: _cuda_devicequery
-
-- name: Set fact for CUDA devices
-  set_fact:
-    cuda_devices: "{{ _cuda_devicequery.stdout | regex_findall('Device (\\d+):') }}"
-
-- name: Run CUDA bandwidth test
-  command:
-    cmd: "{{ cuda_samples_path }}/cuda-samples-{{ cuda_version_short }}/build/Samples/1_Utilities/bandwidthTest/bandwidthTest --device={{ item }}"
-  register: _cuda_bandwidthtest
-  loop: "{{ cuda_devices }}"
-  loop_control:
-    label: "Device {{ item }}" # e.g '0'
-
-- name: Summarise bandwidth test output
-  debug:
-    msg: |
-      {{ _parts[1].splitlines()[0] | trim }}
-      Bandwidths: (Gb/s)
-      Host to Device: {{ _parts[2].split()[-1] }}
-      Device to Host: {{ _parts[3].split()[-1] }}
-      Device to Device: {{ _parts[4].split()[-1] }}
-      {{ ': '.join(_parts[5].split('=') | map('trim')) }}
-      {{ _parts[6] }}
-  loop: "{{ _cuda_bandwidthtest.results }}"
-  vars:
-    _parts: "{{ item.stdout.split('\n\n') }}"
-  loop_control:
-    label: "Device {{ item.item }}" # e.g '0'

ansible/roles/openondemand/README.md

@@ -59,10 +59,10 @@ This role enables SSL on the Open Ondemand server, using the following self-sign
     - `new_window`: Optional. Whether to open link in new window. Bool, default `false`.
     - `app_name`: Optional. Unique name for app appended to `/var/www/ood/apps/sys/`. Default is `name`, useful if that is not unique or not suitable as a path component.
 - `openondemand_dashboard_support_url`: Optional. URL or email etc to show as support contact under Help in dashboard. Default `(undefined)`.
-- `openondemand_desktop_partition`: Optional. Name of Slurm partition to use for remote desktops. Requires a corresponding group named "openondemand_desktop" and entry in openhpc_slurm_partitions.
+- `openondemand_desktop_partition`: Optional. Name of Slurm partition to use for remote desktops. Requires a corresponding group named "openondemand_desktop" and entry in openhpc_partitions.
 - `openondemand_desktop_screensaver`: Optional. Whether to enable screen locking/screensaver. **NB:** Users must have passwords if this is enabled. Bool, default `false`.
 - `openondemand_filesapp_paths`: List of paths (in addition to $HOME, which is always added) to include shortcuts to within the Files dashboard app.
-- `openondemand_jupyter_partition`: Required. Name of Slurm partition to use for Jupyter Notebook servers. Requires a corresponding group named "openondemand_jupyter" and entry in openhpc_slurm_partitions.
+- `openondemand_jupyter_partition`: Required. Name of Slurm partition to use for Jupyter Notebook servers. Requires a corresponding group named "openondemand_jupyter" and entry in openhpc_partitions.
 
 ### Monitoring
 - `openondemand_exporter`: Optional. Install the Prometheus [ondemand_exporter](https://github.com/OSC/ondemand_exporter) on the `openondemand` node to export metrics about Open Ondemand itself. Default `true`.

ansible/roles/rebuild/README.md

@@ -23,7 +23,7 @@ The below are only used by this role's `rebuild.yml` task file, i.e. when
 running the `ansible/adhoc/rebuild-via-slurm.yml` playbook:
 
 - `rebuild_job_partitions`: Optional. Comma-separated list of names of rebuild
-  partitions defined in `openhpc_slurm_partitions`. Useful as an extra-var for
+  partitions defined in `openhpc_partitions`. Useful as an extra-var for
   limiting rebuilds. Default `rebuild`.
 
 - `rebuild_job_name`: Optional. Name of rebuild jobs. Default is `rebuild-`

ansible/validate.yml

@@ -23,12 +23,17 @@
   gather_facts: false
   tags: openhpc
   tasks:
+    - 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/README.md

@@ -0,0 +1,45 @@
+# StackHPC Slurm Appliance Documentation
+
+### Operator docs
+
+[Image build](image-build.md)
+
+[CI](ci.md)
+
+[Monitoring and logging](monitoring-and-logging.md)
+
+[Operations guide](operations.md)
+
+[Production deployment](production.md)
+
+[Upgrades](upgrades.md)
+
+[Sequence diagrams](sequence.md)
+
+### Configuration docs
+
+[Alerting](alerting.md)
+
+[Chrony](chrony.md)
+
+[Environments](environments.md)
+
+[K3s](k3s.README.md)
+
+[Networking](networks.md)
+
+[Open OnDemand](openondemand.md)
+
+[Persistent state](persistent-state.md)
+
+#### Experimental fetaures
+
+[Compute init](experimental/compute-init.md)
+
+[Pulp](experimental/pulp.md)
+
+[Slurm controlled rebuild](experimental/slurm-controlled-rebuild.md)
+
+### Contributor docs
+
+[Adding functionality](adding-functionality.md)

docs/experimental/slurm-controlled-rebuild.md

@@ -107,42 +107,17 @@ The configuration of this is complex and involves:
       defined in the `compute` or `login` variables, to override the default
       image for specific node groups.
 
-5. Modify `openhpc_slurm_partitions` to add a new partition covering rebuildable
-   nodes to use for for rebuild jobs. If using the default OpenTofu
-   configurations, this variable is contained in an OpenTofu-templated file
-   `environments/$ENV/group_vars/all/partitions.yml` which must be overriden
-   by copying it to e.g. a `z_partitions.yml` file in the same directory.
-   However production sites will probably be overriding this file anyway to
-   customise it.
-
-   An example partition definition, given the two node groups "general" and
-   "gpu" shown in Step 2, is:
-
-    ```yaml
-    openhpc_slurm_partitions:
-        ...
-        - name: rebuild
-          groups:
-            - name: general
-            - name: gpu
-          default: NO
-          maxtime: 30
-          partition_params:
-            PriorityJobFactor: 65533
-            Hidden: YES
-            RootOnly: YES
-            DisableRootJobs: NO
-            PreemptMode: 'OFF'
-            OverSubscribe: EXCLUSIVE
-    ```
-
-    Which has parameters as follows:
+5. Ensure `openhpc_partitions` contains a partition covering the nodes to run
+   rebuild jobs. The default definition in `environments/common/inventory/group_vars/all/openhpc.yml`
+   will automatically include this via `openhpc_rebuild_partition` also in that
+   file. If modifying this, note the important parameters are:
+   
     - `name`: Partition name matching `rebuild` role variable `rebuild_partitions`,
       default `rebuild`.
-    - `groups`: A list of node group names, matching keys in the OpenTofu
-      `compute` variable (see example in step 2 above). Normally every compute
-      node group should be listed here, unless Slurm-controlled rebuild is not
-      required for certain node groups.
+    - `groups`: A list of nodegroup names, matching `openhpc_nodegroup` and
+      keys in the OpenTofu `compute` variable (see example in step 2 above).
+      Normally every compute node group should be listed here, unless
+      Slurm-controlled rebuild is not required for certain node groups.
     - `default`: Must be set to `NO` so that it is not the default partition.
     - `maxtime`: Maximum time to allow for rebuild jobs, in
       [slurm.conf format](https://slurm.schedmd.com/slurm.conf.html#OPT_MaxTime).