Leaving default node pool for sanbox to decrease provisioning time

GoogleCloudPlatform · Mar 29, 2024 · bbff3bb · bbff3bb
1 parent cd069f0
commit bbff3bb
Show file tree

Hide file tree

Showing 3 changed files with 25 additions and 2 deletions.
diff --git a/best-practices/ml-platform/examples/platform/sandbox/main.tf b/best-practices/ml-platform/examples/platform/sandbox/main.tf
@@ -156,10 +156,13 @@ module "gke" {
 
   cluster_name                = format("%s-%s", var.cluster_name, var.environment_name)
   env                         = var.environment_name
+  initial_node_count          = 1
+  machine_type                = "n2-standard-8"
   master_auth_networks_ipcidr = var.subnet_01_ip
   network                     = module.create-vpc.vpc
   project_id                  = data.google_project.environment.project_id
   region                      = var.subnet_01_region
+  remove_default_node_pool    = false
   subnet                      = module.create-vpc.subnet-1
   zone                        = "${var.subnet_01_region}-a"
 }

diff --git a/best-practices/ml-platform/terraform/modules/cluster/gke.tf b/best-practices/ml-platform/terraform/modules/cluster/gke.tf
@@ -23,13 +23,13 @@ resource "google_container_cluster" "mlp" {
 
   deletion_protection      = false
   enable_shielded_nodes    = true
-  initial_node_count       = 1
+  initial_node_count       = var.initial_node_count
   location                 = var.region
   name                     = var.cluster_name
   network                  = var.network
   node_locations           = ["${var.region}-a", "${var.region}-b", "${var.region}-c"]
   project                  = var.project_id
-  remove_default_node_pool = true
+  remove_default_node_pool = var.remove_default_node_pool
   subnetwork               = var.subnet
 
   addons_config {
@@ -165,6 +165,8 @@ resource "google_container_cluster" "mlp" {
   }
 
   node_config {
+    machine_type = var.machine_type
+
     shielded_instance_config {
       enable_integrity_monitoring = true
       enable_secure_boot          = true

diff --git a/best-practices/ml-platform/terraform/modules/cluster/variables.tf b/best-practices/ml-platform/terraform/modules/cluster/variables.tf
@@ -23,6 +23,18 @@ variable "env" {
   type        = string
 }
 
+variable "initial_node_count" {
+  default     = 1
+  description = "The number of nodes to create in this cluster's default node pool. In regional or multi-zonal clusters, this is the number of nodes per zone. Must be set if node_pool is not set. If you're using google_container_node_pool objects with no default node pool, you'll need to set this to a value of at least 1, alongside setting remove_default_node_pool to true."
+  type        = number
+}
+
+variable "machine_type" {
+  default     = "e2-medium"
+  description = "The name of a Google Compute Engine machine type."
+  type        = string
+}
+
 variable "master_auth_networks_ipcidr" {
   description = "master authorized network"
   type        = string
@@ -45,6 +57,12 @@ variable "region" {
   type        = string
 }
 
+variable "remove_default_node_pool" {
+  default     = true
+  description = "If true, deletes the default node pool upon cluster creation. If you're using google_container_node_pool resources with no default node pool, this should be set to true, alongside setting initial_node_count to at least 1."
+  type        = bool
+}
+
 variable "subnet" {
   description = "subnetwork where the cluster will be created"
   type        = string