Merge pull request #27 from NVIDIA/magzhang/0.7.0-updates

0.7.0 - Updated to k8s 1.29 and GPU Operator v23.9.2
NVIDIA · Apr 18, 2024 · 8977e19 · 8977e19
2 parents b388bcd + 59b7fcd
commit 8977e19
Show file tree

Hide file tree

Showing 8 changed files with 33 additions and 30 deletions.
diff --git a/README.md b/README.md
@@ -24,6 +24,7 @@ Each CSP has its own end of life date for the versions of Kubernetes they suppor
 
 | Version | Release Date  | Kubernetes Versions                            | NVIDIA GPU Operator    | NVIDIA Data Center Driver* | End of Life |
 | :---    |    :---       | :---                                           | :---                   | :---                      | :--- |
+| 0.7.0     | April 2024 | EKS -  1.29 <br> GKE - 1.29 <br> AKS - 1.29 | 23.9.2 (Default); 23.9.2 (NV AI E)                 | 550.54.15  (EKS & GKE Default); 550.54.15 (NV AI E version for GKE & EKS)    | EKS - Mar 2025 <br> GKE - Mar 2025  <br> AKS - Not Specified |
 | 0.6.0     | January 2024 | EKS -  1.28 <br> GKE - 1.28 <br> AKS - 1.28 | 23.9.1 (Default); 23.9.0 (NV AI E)                 | 535.129.03  (EKS & GKE Default); 535.129.03 (NV AI E version for GKE & EKS)    | EKS - Nov 2024 <br> GKE - Nov 2024  <br> AKS - Nov 2024 |
 | 0.5.0     | November 2023 | EKS -  1.27 <br> GKE - 1.27 <br> AKS - 1.27 | 23.6.1 (Default); 23.3.2 (NV AI E)                 | 535.104.05  (EKS & GKE Default); 525.125.06 (NV AI E version for GKE & EKS)    | EKS - July 2024 <br> GKE - August 2024  <br> AKS - July 2024 |
 | 0.4.0     | October 2023 | EKS -  1.27 <br> GKE - 1.27 <br> AKS - 1.27 | 23.6.1 (Default); 23.3.2 (NV AI E)                 | 535.104.05  (EKS & GKE Default); 525.125.06 (NV AI E version for GKE & EKS)    | EKS - July 2024 <br> GKE - August 2024  <br> AKS - July 2024 |

diff --git a/aks/examples/cnpack/README.md b/aks/examples/cnpack/README.md
@@ -24,13 +24,15 @@
     - Add `fluentbit_workspace_name`. This will create Azure Log Analytics Workspace with the specified name.
     - Add `prometheus_name`. This will create Azure Monitor Workspace with the specified name.
 
-2. Run `terraform plan -out tfplan` and validate that the output is correct
+2. Run `terraform init`
 
-3. Run `terraform apply tfplan`
+3. Run `terraform plan -out tfplan` and validate that the output is correct
 
-4. The `terraform output` of this module can be used immediately within the configuration file of CNPack
+4. Run `terraform apply tfplan`
 
-5. Run `terraform destroy` to delete all resources created by this module.
+5. The `terraform output` of this module can be used immediately within the configuration file of CNPack
+
+6. Once you're done, run `terraform state rm module.holoscan-ready-aks.helm_release.gpu-operator` and `terraform state rm module.holoscan-ready-aks.kubernetes_namespace_v1.gpu-operator`. Lastly, run `terraform destroy` to delete all resources created by this module.
 
 **Note**
 The `log_analytics_workspace_primary_shared_key` used for Fluentbit is a sensitive variable and should be protected like a password

diff --git a/aks/terraform.tfvars b/aks/terraform.tfvars
@@ -19,9 +19,9 @@
 # gpu_node_pool_max_count      = 5
 # gpu_node_pool_min_count      = 2
 # gpu_operator_namespace       = "gpu-operator"
-# gpu_operator_version         = "v23.9.1"
+# gpu_operator_version         = "v23.9.2"
 # gpu_os_sku                   = "Ubuntu"
-# kubernetes_version           = "1.28"
+# kubernetes_version           = "1.29"
 # location                     = ""
 # nvaie                        = false
-# nvaie_gpu_operator_version   = "v23.9.0"
+# nvaie_gpu_operator_version   = "v23.9.2"
diff --git a/aks/variables.tf b/aks/variables.tf
@@ -25,7 +25,7 @@ variable "cluster_name" {
 }
 
 variable "kubernetes_version" {
-  default     = "1.28"
+  default     = "1.29"
   description = "Version of Kubernetes to turn on. Run 'az aks get-versions --location <location> --output table' to view all available versions "
 }
 
@@ -87,7 +87,7 @@ variable "gpu_os_sku" {
 GPU Operator Variables
 ****************************/
 variable "gpu_operator_version" {
-  default     = "v23.9.1"
+  default     = "v23.9.2"
   description = "Version of the GPU operator to be installed"
 }
 
@@ -105,7 +105,7 @@ variable "nvaie" {
 
 variable "nvaie_gpu_operator_version" {
   type        = string
-  default     = "v23.9.0"
+  default     = "v23.9.2"
   description = "The NVIDIA Driver version of GPU Operator. Overrides `gpu_operator_version` when `nvaie` is set to `true`"
 }
 

diff --git a/eks/terraform.tfvars b/eks/terraform.tfvars
@@ -10,7 +10,7 @@
 # aws_profile                           = "development"
 # cidr_block                            = "10.0.0.0/16"
 # cluster_name                          = ""
-# cluster_version                       = "1.28"
+# cluster_version                       = "1.29"
 # cpu_instance_type                     = "t2.xlarge"
 # cpu_node_pool_additional_user_data    = ""
 # cpu_node_pool_delete_on_termination   = true
@@ -28,16 +28,16 @@
 # gpu_node_pool_delete_on_termination   = true
 # gpu_node_pool_root_disk_size_gb       = 512
 # gpu_node_pool_root_volume_type        = "gp2"
-# gpu_operator_driver_version           = "535.129.03"
+# gpu_operator_driver_version           = "550.54.15"
 # gpu_operator_namespace                = "gpu-operator"
-# gpu_operator_version                  = "v23.9.1"
+# gpu_operator_version                  = "v23.9.2"
 # max_cpu_nodes                         = "2"
 # max_gpu_nodes                         = "5"
 # min_cpu_nodes                         = "0"
 # min_gpu_nodes                         = "2"
 # nvaie                                 = false
-# nvaie_gpu_operator_driver_version     = "535.129.03"
-# nvaie_gpu_operator_version            = "v23.9.0"
+# nvaie_gpu_operator_driver_version     = "550.54.15"
+# nvaie_gpu_operator_version            = "v23.9.2"
 # private_subnets = [
 #   "10.0.0.0/19",
 #   "10.0.32.0/19",

diff --git a/eks/variables.tf b/eks/variables.tf
@@ -28,20 +28,20 @@ variable "cluster_name" {
 
 variable "cluster_version" {
   type        = string
-  default     = "1.28"
+  default     = "1.29"
   description = "Version of EKS to install on the control plane (Major and Minor version only, do not include the patch)"
 }
 /************************
   GPU Operator Variables
 *************************/
 variable "gpu_operator_version" {
-  default     = "v23.9.1"
+  default     = "v23.9.2"
   description = "Version of the GPU Operator to deploy. Defaults to latest available. Not set when `nvaie` is set to `true`"
 }
 
 variable "gpu_operator_driver_version" {
   type        = string
-  default     = "535.129.03"
+  default     = "550.54.15"
   description = "The NVIDIA Driver version deployed with GPU Operator. Defaults to latest available. Not set when `nvaie` is set to true"
 }
 
@@ -59,13 +59,13 @@ variable "nvaie" {
 
 variable "nvaie_gpu_operator_version" {
   type        = string
-  default     = "v23.9.0"
+  default     = "v23.9.2"
   description = "The NVIDIA Driver version of GPU Operator. Overrides `gpu_operator_version` when `nvaie` is set to `true`"
 }
 
 variable "nvaie_gpu_operator_driver_version" {
   type        = string
-  default     = "535.129.03"
+  default     = "550.54.15"
   description = "The NVIDIA AI Enterprise version of the NVIDIA driver to be installed with the GPU operator. Overrides `gpu_operator_driver_version` when `nvaie` is set to `true`"
 }
 /*****************************

diff --git a/gke/terraform.tfvars b/gke/terraform.tfvars
@@ -14,18 +14,18 @@
 # gpu_instance_type                 = "n1-standard-4"
 # gpu_max_node_count                = "5"
 # gpu_min_node_count                = "2"
-# gpu_operator_driver_version       = "535.129.03"
+# gpu_operator_driver_version       = "550.54.15"
 # gpu_operator_namespace            = "gpu-operator"
-# gpu_operator_version              = "v23.9.1"
+# gpu_operator_version              = "v23.9.2"
 # gpu_type                          = "nvidia-tesla-v100"
-# min_master_version                = "1.28"
+# min_master_version                = "1.29"
 # network                           = ""
 # node_zones                        = ""
 # num_cpu_nodes                     = 1
 # num_gpu_nodes                     = 2
 # nvaie                             = false
-# nvaie_gpu_operator_driver_version = "535.129.03"
-# nvaie_gpu_operator_version        = "v23.9.0"
+# nvaie_gpu_operator_driver_version = "550.54.15"
+# nvaie_gpu_operator_version        = "v23.9.2"
 # project_id                        = ""
 # region                            = ""
 # release_channel                   = "REGULAR"

diff --git a/gke/variables.tf b/gke/variables.tf
@@ -49,7 +49,7 @@ variable "release_channel" {
 }
 
 variable "min_master_version" {
-  default     = "1.28"
+  default     = "1.29"
   description = "The minimum cluster version of the master."
 }
 
@@ -133,13 +133,13 @@ variable "disk_size_gb" {
 GPU Operator Variables
 ***************************/
 variable "gpu_operator_version" {
-  default     = "v23.9.1"
+  default     = "v23.9.2"
   description = "Version of the GPU Operator to deploy. Defaults to latest available. Not set when `nvaie` is set to `true`"
 }
 
 variable "gpu_operator_driver_version" {
   type        = string
-  default     = "535.129.03"
+  default     = "550.54.15"
   description = "The NVIDIA Driver version deployed with GPU Operator. Defaults to latest available. Not set when `nvaie` is set to true"
 }
 
@@ -157,12 +157,12 @@ variable "nvaie" {
 
 variable "nvaie_gpu_operator_version" {
   type        = string
-  default     = "v23.9.0"
+  default     = "v23.9.2"
   description = "The NVIDIA Driver version of GPU Operator. Overrides `gpu_operator_version` when `nvaie` is set to `true`"
 }
 
 variable "nvaie_gpu_operator_driver_version" {
   type        = string
-  default     = "535.129.03"
+  default     = "550.54.15"
   description = "The NVIDIA AI Enterprise version of the NVIDIA driver to be installed with the GPU operator. Overrides `gpu_operator_driver_version` when `nvaie` is set to `true`"
 }