Merge branch 'master' into feat/b64-mungekey

Author: sjpb

.github/workflows/ci.yml

@@ -14,7 +14,7 @@ jobs:
     # as the container command is flaky.
     # This job builds an image using the upstream rockylinux:9.3 image which ensures
     # that the image used for the molecule workflow is always updated.
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
     defaults:
       run:
         working-directory: molecule/images
@@ -36,7 +36,7 @@ jobs:
 
   molecule:
     name: Molecule
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
     needs: build
     strategy:
       fail-fast: false
@@ -108,7 +108,7 @@ jobs:
 
   checks:
     name: Checks
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
     steps:
       - name: Check out the codebase.
         uses: actions/checkout@v3

README.md

@@ -107,7 +107,14 @@ partition configuration for each.
 [slurm.conf](https://slurm.schedmd.com/slurm.conf.html). Keys are slurm.conf
 parameter names and values are lists or strings as appropriate. This can be
 used to supplement or override the template defaults. Templated parameters can
-also be removed by setting the value to the literal string`'omit'` - note
+also be removed by setting the value to the literal string `'omit'` - note
+that this is *not the same* as the Ansible `omit` [special variable](https://docs.ansible.com/ansible/latest/reference_appendices/special_variables.html#term-omit).
+
+`openhpc_cgroup_config`: Optional. Mapping of additional parameters and values for
+[cgroup.conf](https://slurm.schedmd.com/cgroup.conf.html). Keys are cgroup.conf
+parameter names and values are lists or strings as appropriate. This can be
+used to supplement or override the template defaults. Templated parameters can
+also be removed by setting the value to the literal string `'omit'` - note
 that this is *not the same* as the Ansible `omit` [special variable](https://docs.ansible.com/ansible/latest/reference_appendices/special_variables.html#term-omit).
 
 `openhpc_ram_multiplier`: Optional, default `0.95`. Multiplier used in the calculation: `total_memory * openhpc_ram_multiplier` when setting `RealMemory` for the partition in slurm.conf. Can be overriden on a per partition basis using `openhpc_slurm_partitions.ram_multiplier`. Has no effect if `openhpc_slurm_partitions.ram_mb` is set.

defaults/main.yml

@@ -40,8 +40,14 @@ openhpc_default_config:
   PropagateResourceLimitsExcept: MEMLOCK
   Epilog: /etc/slurm/slurm.epilog.clean
   ReturnToService: 2
+openhpc_cgroup_default_config:
+  ConstrainCores: "yes"
+  ConstrainDevices: "yes"
+  ConstrainRAMSpace: "yes"
+  ConstrainSwapSpace: "yes"
 
 openhpc_config: {}
+openhpc_cgroup_config: {}
 openhpc_gres_template: gres.conf.j2
 openhpc_slurm_configless: "{{ 'enable_configless' in openhpc_config.get('SlurmctldParameters', []) }}"
 

files/nodegroup.schema

@@ -0,0 +1,86 @@
+{  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "type": "object",
+  "definitions": {
+    "gres": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "properties": {
+          "conf": {
+            "type": "string",
+            "minLength": 1
+          },
+          "file": {
+            "type": "string",
+            "minLength": 1
+          }
+        },
+        "required": [
+          "conf"
+        ]
+      }
+    }
+  },
+  "properties": {
+    "name": {
+      "type": "string",
+      "minLength": 1
+    },
+    "ram_mb": {
+      "type": "number",
+    },
+    "ram_multiplier": {
+      "type": "number",
+    },
+    "features": {
+    	"type": "array",
+        "items": {
+        	"type": "string"
+        }
+    },
+    "node_params": {
+    	"type": "object",
+    },            
+    "gres_autodetect": {
+      "type": "string",
+      "minLength": 1
+    },
+    "gres": {
+      "$ref": "#/definitions/gres"
+    }
+  },
+  "required": [
+    "name"
+  ],
+  "if": {
+    "properties": {
+      "gres_autodetect": {
+        "const": "off"
+      }
+    }
+  },
+  "then": {
+    "properties": {
+      "gres": {
+        "items": {
+          "required": [
+            "file"
+          ]
+        }
+      }
+    }
+  },
+  "else": {
+    "properties": {
+      "gres": {
+        "items": {
+          "not": {
+            "required": [
+              "file"
+            ]
+          }
+        }
+      }
+    }
+  }
+}

tasks/runtime.yml

@@ -91,6 +91,10 @@
     owner: root
     group: root
   when: openhpc_enable.control | default(false) or not openhpc_slurm_configless | bool
+  notify:
+    - Restart slurmctld service
+  register: ohpc_cgroup_conf
+  # NB uses restart rather than reload as this is needed in some cases
 
 - name: Remove local tempfile for slurm.conf templating
   ansible.builtin.file:
@@ -125,7 +129,7 @@
   changed_when: true
   when:
     - openhpc_slurm_control_host in ansible_play_hosts
-    - hostvars[openhpc_slurm_control_host].ohpc_slurm_conf.changed or hostvars[openhpc_slurm_control_host].ohpc_gres_conf.changed # noqa no-handler
+    - hostvars[openhpc_slurm_control_host].ohpc_slurm_conf.changed or hostvars[openhpc_slurm_control_host].ohpc_cgroup_conf.changed or hostvars[openhpc_slurm_control_host].ohpc_gres_conf.changed # noqa no-handler
   notify:
     - Restart slurmd service
 

tasks/validate.yml

@@ -21,23 +21,15 @@
   delegate_to: localhost
   run_once: true
 
-- name: Check gres configuration when gres_autodetect is set
-  assert:
-    that:
-      - _failure_reasons | selectattr('when', 'truthy') | length == 0
-    fail_msg: >
-      Your nodegroup definition must include a single gres dictionary containing a conf key
-      if gres_autodetect is set. The following nodegroup failed this check: {{ item }}.
-      Reasons for failure: {{ _failure_reasons | selectattr('when', 'truthy') | map(attribute='msg') | join(', ') }}
+- name: Validate openhpc_nodegroups
+  ansible.utils.validate:
+    criteria: "{{ lookup('file', 'nodegroup.schema') }}"
+    engine: 'ansible.utils.jsonschema'
+    data: "{{ item }}"
   vars:
-    _openhpc_gres_autodetect_groups: "{{ openhpc_nodegroups | selectattr('gres_autodetect', 'defined') | selectattr('gres_autodetect', 'search', '(?!off).*') }}"
-    _failure_reasons:
-      - msg: The gres key was a list with more than one item
-        when: "{{ item.gres | length != 1 }}"
-      - msg: The gres dictionary does not contain a conf key
-        when: "{{ item.gres.0.conf is not defined }}"
+    ansible_jsonschema_draft: '2020-12'
   delegate_to: localhost
-  loop: "{{ _openhpc_gres_autodetect_groups }}"
+  loop: "{{ openhpc_nodegroups }}"
   run_once: true
 
 - name: Fail if partition configuration is outdated

templates/cgroup.conf.j2

@@ -5,7 +5,8 @@
 # See man slurm.conf and man cgroup.conf for further
 # information on cgroup configuration parameters
 #--
-ConstrainCores=yes
-ConstrainDevices=yes
-ConstrainRAMSpace=yes
-ConstrainSwapSpace=yes
+{% for k, v in openhpc_cgroup_default_config | combine(openhpc_cgroup_config) | items %}
+{%  if v != "omit" %}{# allow removing items using setting key: null #}
+{{ k }}={{ v | join(',') if (v is sequence and v is not string) else v }}
+{%  endif %}
+{% endfor %}