viktor/infra
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1

set the time slicing config in the nvidia chart values[ci skip]

Browse source
This commit is contained in:
Viktor Barzin 2025-12-28 08:35:36 +00:00
parent 10ab0d7c40
commit 8af9e6b5bd
3 changed files with 10 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

1
modules/kubernetes/authentik/pgbouncer.tf
View file

@ -93,6 +93,7 @@ resource "kubernetes_deployment" "pgbouncer" {
} }
} }
} }
depends_on = [kubernetes_secret.pgbouncer_auth]
} }
# --- 4️⃣ Service --- # --- 4️⃣ Service ---

10
modules/kubernetes/nvidia/main.tf
View file

@ -15,7 +15,7 @@ resource "kubernetes_namespace" "nvidia" {
} }
} }
# Apply to operator with: # [not needed anymore; part of the chart values] Apply to operator with:
# kubectl patch clusterpolicies.nvidia.com/cluster-policy -n gpu-operator --type merge -p '{"spec": {"devicePlugin": {"config": {"name": "time-slicing-config", "default": "any"}}}}' # kubectl patch clusterpolicies.nvidia.com/cluster-policy -n gpu-operator --type merge -p '{"spec": {"devicePlugin": {"config": {"name": "time-slicing-config", "default": "any"}}}}'
resource "kubernetes_config_map" "time_slicing_config" { resource "kubernetes_config_map" "time_slicing_config" {
@ -37,6 +37,7 @@ resource "kubernetes_config_map" "time_slicing_config" {
replicas: 10 replicas: 10
EOF EOF
} }
depends_on = [kubernetes_namespace.nvidia]
} }
resource "helm_release" "nvidia-gpu-operator" { resource "helm_release" "nvidia-gpu-operator" {
@ -49,7 +50,8 @@ resource "helm_release" "nvidia-gpu-operator" {
# version = "0.9.3" # version = "0.9.3"
timeout = 6000 timeout = 6000
values = [templatefile("${path.module}/values.yaml", {})] values = [templatefile("${path.module}/values.yaml", {})]
depends_on = [kubernetes_config_map.time_slicing_config]
} }
resource "kubernetes_deployment" "nvidia-exporter" { resource "kubernetes_deployment" "nvidia-exporter" {
@ -83,9 +85,6 @@ resource "kubernetes_deployment" "nvidia-exporter" {
port { port {
container_port = 9400 container_port = 9400
} }
port {
container_port = 9400
}
security_context { security_context {
privileged = true privileged = true
capabilities { capabilities {
@ -101,6 +100,7 @@ resource "kubernetes_deployment" "nvidia-exporter" {
} }
} }
} }
depends_on = [helm_release.nvidia-gpu-operator]
} }
resource "kubernetes_service" "nvidia-exporter" { resource "kubernetes_service" "nvidia-exporter" {

4
modules/kubernetes/nvidia/values.yaml
View file

@ -13,3 +13,7 @@ driver:
version: "570.195.03" # CUDA 12.8 version: "570.195.03" # CUDA 12.8
upgradePolicy: upgradePolicy:
autoUpgrade: false autoUpgrade: false
devicePlugin:
config:
name: time-slicing-config