Revert "Add support for configuring Multi-Instance GPUs (MIG) (#656)"

This reverts commit 7509986.

Author: sjpb

.github/workflows/extra.yml

@@ -32,11 +32,11 @@ jobs:
           - image_name: openhpc-extra-RL8
             source_image_name_key: RL8 # key into environments/.stackhpc/tofu/cluster_image.auto.tfvars.json
             inventory_groups: doca,cuda,lustre
-            volume_size: 35 # needed for cuda
+            volume_size: 30 # needed for cuda
           - image_name: openhpc-extra-RL9
             source_image_name_key: RL9
             inventory_groups: doca,cuda,lustre
-            volume_size: 35 # needed for cuda
+            volume_size: 30 # needed for cuda
     env:
       ANSIBLE_FORCE_COLOR: True
       OS_CLOUD: openstack

ansible/.gitignore

@@ -90,7 +90,5 @@ roles/*
 !roles/gateway/**
 !roles/alertmanager/
 !roles/alertmanager/**
-!roles/slurm_recompile/**
-!roles/slurm_recompile/**
 !roles/nhc/
 !roles/nhc/**

ansible/extras.yml

@@ -48,20 +48,6 @@
         name: cuda
         tasks_from: "{{ 'runtime.yml' if appliances_mode == 'configure' else 'install.yml' }}"
 
-- name: Setup vGPU
-  hosts: vgpu
-  become: yes
-  gather_facts: yes
-  tags: vgpu
-  tasks:
-    - include_role:
-        name: stackhpc.linux.vgpu
-        tasks_from: "{{ 'configure.yml' if appliances_mode == 'configure' else 'install.yml' }}"
-  handlers:
-    - name: reboot
-      fail:
-        msg: Reboot handler for stackhpc.linux.vgpu role fired unexpectedly. This was supposed to be unreachable.
-
 - name: Persist hostkeys across rebuilds
   # Must be after filesystems.yml (for storage)
   # and before portal.yml (where OOD login node hostkeys are scanned)

ansible/fatimage.yml

@@ -250,16 +250,6 @@
         name: cloudalchemy.grafana
         tasks_from: install.yml
 
-- name: Add support for NVIDIA GPU auto detection to Slurm
-  hosts: cuda
-  become: yes
-  tasks:
-    - name: Recompile slurm
-      import_role:
-        name: slurm_recompile
-      vars:
-        slurm_recompile_with_nvml: "{{ groups.cuda | length > 0 }}"
-
 - name: Run post.yml hook
   vars:
     appliances_environment_root: "{{ lookup('env', 'APPLIANCES_ENVIRONMENT_ROOT') }}"

ansible/roles/compute_init/README.md

@@ -75,7 +75,6 @@ it also requires an image build with the role name added to the
 | extras.yml               | basic_users             | All functionality [6]           | No                  |
 | extras.yml               | eessi                   | All functionality [7]           | No                  |
 | extras.yml               | cuda                    | None required - use image build | Yes [8]             |
-| extras.yml               | vgpu                    | All functionality               | Yes                 |
 | extras.yml               | persist_hostkeys        | Not relevant for compute nodes  | n/a                 |
 | extras.yml               | compute_init (export)   | Not relevant for compute nodes  | n/a                 |
 | extras.yml               | k9s (install)           | Not relevant during boot        | n/a                 |

ansible/roles/compute_init/files/compute-init.yml

@@ -19,7 +19,6 @@
     enable_basic_users: "{{ os_metadata.meta.basic_users | default(false) | bool }}"
     enable_eessi: "{{ os_metadata.meta.eessi | default(false) | bool }}"
     enable_chrony: "{{ os_metadata.meta.chrony | default(false) | bool }}"
-    enable_vgpu: "{{ os_metadata.meta.vpgu | default(false) | bool }}"
     enable_nhc: "{{ os_metadata.meta.nhc | default(false) | bool }}"
 
     # TODO: "= role defaults" - could be moved to a vars_file: on play with similar precedence effects
@@ -297,12 +296,6 @@
             cmd: "cvmfs_config setup"
       when: enable_eessi
 
-    - name: Configure VGPUs
-      include_role:
-        name: stackhpc.linux.vgpu
-        tasks_from: 'configure.yml'
-      when: enable_vgpu
-
     # NB: don't need conditional block on enable_compute as have already exited
     # if not the case
     - name: Write Munge key