Fix tuned hpc-compute with hugepages and verify applied profile

ansible/roles/tuned/tasks/configure.yml

@@ -12,9 +12,17 @@
   register: _tuned_profile_current
   changed_when: false
 
-- name: Set tuned-adm profile
+- name: Set TuneD profile
   ansible.builtin.command:
     cmd: "tuned-adm profile {{ tuned_profile }}"
-  when: 
+  when:
+    - tuned_started | bool
+    - tuned_profile not in _tuned_profile_current.stdout
+
+- name: Verify TuneD profile
+  ansible.builtin.command:
+    cmd: tuned-adm verify
+  changed_when: false
+  when:
     - tuned_started | bool
     - tuned_profile not in _tuned_profile_current.stdout

ansible/roles/tuned/tasks/install.yml

@@ -2,4 +2,15 @@
 - name: Install tuneD
   ansible.builtin.dnf:
     name: tuned
-    state: present
+    state: present
+
+- name: Fix TuneD hpc-compute profile for hugepages
+  # See https://github.com/redhat-performance/tuned/issues/752
+  # This is done on install, not configure, so that it is available even
+  # for compute-init nodes
+  community.general.ini_file:
+    path: /usr/lib/tuned/hpc-compute/tuned.conf
+    section: sysctl
+    option: vm.min_free_kbytes
+    value: '>135168'
+    no_extra_spaces: true

environments/.stackhpc/inventory/extra_groups

@@ -29,7 +29,10 @@ cluster
 
 [tuned:children]
 # Install tuned into fat image
+# NB: builder has tuned_enabled and tuned_started false so does not configure it
 builder
+# Also test tuned during site playbook
+cluster
 
 [squid:children]
 # Install squid into fat image

environments/.stackhpc/inventory/group_vars/all/tuned.yml

@@ -0,0 +1,2 @@
+# Set profile which is not default (on VMs) for testing
+tuned_profile: hpc-compute

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

@@ -1,6 +1,6 @@
 {
     "cluster_image": {
-        "RL8": "openhpc-RL8-250514-1502-5a923b2c",
-        "RL9": "openhpc-RL9-250514-1502-5a923b2c"
+        "RL8": "openhpc-RL8-250610-1435-d0ef926e",
+        "RL9": "openhpc-RL9-250610-1435-d0ef926e"
     }
 }

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

@@ -36,7 +36,7 @@ nfs_configurations_extra: [] # site-specific nfs shares
 
 nfs_configurations: >- # construct stackhpc.nfs variable
   {{
-    nfs_configuration_home_volume if (cluster_home_volume | default(true)) else []
+    (nfs_configuration_home_volume if (cluster_home_volume | default(true)) else [])
     +
     nfs_configuration_compute_nodes
     +