From 341094766cd3908cdd6620ba591749e8772ec515 Mon Sep 17 00:00:00 2001
From: Steve Brasier <33413598+sjpb@users.noreply.github.com>
Date: Thu, 8 May 2025 10:30:15 +0100
Subject: [PATCH] Update openhpc.yml

Fix some issues with partition config:

- "Default" partition missing partition_params and default. Although actually, I suspect this is no matching inventory group and hence no nodes in it - so probably the entire partition is irrelevant and shouldl be removed?
- default values should be strings, bit worried bare NO/YES do some horrible ansible autoconversion to bool
- made partition params case consistent, although actually SLurm won't care.
---
 .../vtest/inventory/group_vars/all/openhpc.yml     | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/environments/vtest/inventory/group_vars/all/openhpc.yml b/environments/vtest/inventory/group_vars/all/openhpc.yml
index eb59677ab..b1d5aeef5 100644
--- a/environments/vtest/inventory/group_vars/all/openhpc.yml
+++ b/environments/vtest/inventory/group_vars/all/openhpc.yml
@@ -18,9 +18,11 @@ openhpc_packages_extra: "{{ openhpc_packages_extra_nrel }}"
 
 openhpc_slurm_partitions:
   - name: "DEFAULT"
-    PreemptMode: "OFF"
+    default: "NO"
+    partition_params:
+      PreemptMode: "OFF"
   - name: "gpu"
-    default: NO
+    default: "NO"
     maxtime: "2-0" # 2 days 0 hours
     groups:
       - name: "gpu0"
@@ -34,13 +36,13 @@ openhpc_slurm_partitions:
           Features: "gpu,a100,a100-40"
         gres:
           - conf: "gpu:a100:4"
-            file: "/dev/nvidia[0-3]" 
+            file: "/dev/nvidia[0-3]"
       - name: "gpu3"
         node_params:
           Features: "gpu,a100,a100-80"
         gres:
           - conf: "gpu:a100:8"
-            file: "/dev/nvidia[0-7]" 
+            file: "/dev/nvidia[0-7]"
     partition_params:
       DefMemPerCPU: "1024"
       QOS: "p_gpu"
@@ -49,10 +51,10 @@ openhpc_slurm_partitions:
         #- conf: "gpu:a100:1"
         #file: "/dev/nvidia0"
   - name: "lg"
-    default: YES
+    default: "YES"
     maxtime: "2-0" # 2 days 0 hours
     partition_params:
-      qos: "p_lg"
+      QOS: "p_lg"
       OverSubscribe: "no"
       DefMemPerCPU: "1024"