Skip to content

Latest commit

 

History

History
135 lines (79 loc) · 5.31 KB

tutorial.md

File metadata and controls

135 lines (79 loc) · 5.31 KB
Raw

Let's get started!

Welcome to the Cloudshell tutorial for AI on GKE!

This guide will show you how to prepare GKE cluster and install the AI applications on GKE. It'll also walk you through the files that needs your inputs and commands that will complete the tutorial.

Add more summary here

Time to complete: About 20 minutes

Prerequisites: A Cloud Billing account

Click the Continue button to move to the next step.

What is AI-on-GKE

The AI-on-GKE provides Lorem Ipsum is simply dummy text of the printing and typesetting industry. Lorem Ipsum has been the industry's standard dummy text ever since the 1500s, when an unknown printer took a galley of type and scrambled it to make a type specimen book. It has survived not only five centuries.

Next, you'll provide inputs parameters and launch a AI-on-GKE tutorial.

Step 1: Provide PLATFORM Inputs Parameters for Terraform

Here on step 1 you need to update the PLATFORM terraform tfvars file (located in ./platform/platform.auto.tfvars) to provide the input parameters to allow terraform code execution to provision GKE resource. Thise will include the inputs parameters in form of key value pairs. Update the values as per your requirements.

Open platform.auto.tfvars

Tip: Click the highlighted text above to open the file on your cloudshell.

Step 2: Provide APPLICATION Inputs Parameters for Terraform

Here on step 1 you need to update the APPLICATION terraform tfvars file (located in ./workloads/workloads.auto.tfvars) to provide the input parameters to allow terraform code execution to provision the APPLICATION WORKLOADS. Thise will include the inputs parameters in form of key value pairs. Update the values as per your requirements.

Open workloads.auto.tfvars

Tip: Click the highlighted text above to open the file on your cloudshell.

Step 3: [Optional] Configure Terraform GCS backend

You can also configure the GCS bucket to persist the terraform state file for further usage. To configure the terraform backend you need to have GCS bucket already created. This needs to be done both for PLATFORM and APPLICATION stages.

[Optional] Create GCS Bucket

In case you dont have GCS bucket already, you can create using terraform or gcloud command as well. Refer below for gcloud command line to create new GCS bucket.

gcloud storage buckets create gs://BUCKET_NAME

Tip: Click the copy button on the side of the code box to paste the command in the Cloud Shell terminal to run it.

[Optional] Modify PLATFORM Terraform State Backend

Modify the ./platform/backend.tf and uncomment the code and update the backend bucket name. Open ./platform/backend.tf

After changes file will look like below:

terraform {
 backend "gcs" {
   bucket  = "BUCKET_NAME"
   prefix  = "terraform/state"
 }
}

Refer here for more details.

[Optional] APPLICATION Terraform State Backend

Modify the ./workloads/backend.tf and uncomment the code and update the backend bucket name. Open ./workloads/backend.tf

After changes file will look like below:

terraform {
 backend "gcs" {
   bucket  = "BUCKET_NAME"
   prefix  = "terraform/state/workloads"
 }
}

Refer here for more details.

Step 4: Configure Cloud Build Service Account

Modify the ./cloudbuild.yaml file to provide an existing Service Account with permissions to deploy the infrastructure and workloads components. This service account is used by the Cloud Build service to create the environment.

Open ./cloudbuild.yaml

Step 5: Run Terraform Apply using Cloud Build

You are ready to deploy your resources now! cloudbuild.yaml is already prepared with all the steps requires to deploy the application.

To set your Cloud Platform project in this terminal session use:

gcloud config set project [PROJECT_ID]

Run the below command to submit cloudbuild job to deploy the resources:

## Deploy Private GKE with Connect Gateway and deploy workloads
gcloud beta builds submit --config=cloudbuild.yaml --substitutions=_PLATFORM_VAR_FILE="private-standard-gke-with-new-network.platform.tfvars",_WORKLOADS_VAR_FILE="private-gke-workloads.tfvars"


## Deploy Public GKE with Authorised Networks and deploy workloads
gcloud beta builds submit --config=cloudbuild.yaml --substitutions=_PLATFORM_VAR_FILE="public-standard-gke-with-new-network.platform.tfvars",_WORKLOADS_VAR_FILE="public-gke-workloads.tfvars"

Monitor the terminal for the log link and status for cloud build jobs.

Congratulations

You're all set!

You can now access your cluster and applications.