Demo Alibaba Workshop

Date: 10-03-2023 Video: Onedrive Slide: Slide

Target

• Quickly spin up infrastructure for a MVP node js application
• Setup CICD to automate the deployment from github to ACK

Folder structure

├── Dockerfile              
├── Makefile                # shorthand commands
├── README.md
├── index.js
├── .github                 # Github action workflow
└── terraform               # IaC code
    ├── cr.tf
    ├── main.tf
    ├── provider.tf
    ├── terraform.tfvars
    └── variables.tf

Instruction

1. Create a RAM user and download the access key. For this demo, we can use AdministratorAccess

2. Export credential env to avoid hard code to the terraform code. In this demo, we use ap-southeast-1 (Singapore) as the main region to provision our resources.

export ALICLOUD_ACCESS_KEY=YOUR_KEY
export ALICLOUD_SECRET_KEY=YOUR_SECRET
export ALICLOUD_REGION=ap-southeast-1

3. Clone the code to your folder.

git clone https://github.com/owl-king/demo-workshop-10032023

4. Provision our infrastructure by terraform. You can find the installation instruction in the link

cd demo-workshop-10032023/terraform
terraform init

# Check the current setup. Modify it if you need to change some configs
terraform plan # See what changes are going to apply to our infrastructure
terraform apply 
# around 7m to create a k8s cluster
# 1-2m to creata a node pool

5. Authenticate to k8s cluster Simple way:

• Open alibaba console => Container service => Select our cluster hello-mvp.
• Select Connection Information tab => Generate Temporary kubeconfig
• Create config file at ~/.kube/config and paste the content from temporary kubeconfig

Verify the connection by kubectl cluster-info

6. Login to Docker registry

• Obtain credential by visiting registry-credential
• Login from commandline

sudo docker login --username=YOUR_USER registry-intl.ap-southeast-1.aliyuncs.com

7. Build and push the image

# Get short commit sha
TAG=`git rev-parse --short HEAD`
IMAGE="registry-intl.ap-southeast-1.aliyuncs.com/hello-mvp/hello-nodejs"

# For m1 mac
make build-image-m1 IMAGE=$IMAGE TAG=$TAG

# For intel
make build-image IMAGE=$IMAGE TAG=$TAG

# Push image

make push-image IMAGE=$IMAGE TAG=$TAG

8. Make our service online

TAG=`git rev-parse --short HEAD`
IMAGE="registry-intl.ap-southeast-1.aliyuncs.com/hello-mvp/hello-nodejs"
DEPLOYMENT_NAME=hello
SERVICE_NAME=hello

make create-deployment DEPLOYMENT_NAME=$DEPLOYMENT_NAME IMAGE=$IMAGE TAG=$TAG
make create-service SERVICE_NAME=$SERVICE_NAME


# See all the resources we have created
kubectl get all

# Watch the pod status
kubectl get pod -w

# Check the service status 
kubectl get svc $SERVICE_NAME

9. Setup our cicd pipeline With the current workflow, it will automatically triggered whenever the code is deployed to master. Current workflow has several steps:

• calculate the short sha
• login to docker registry
• build and push the image
• grab the ack credential
• update the deployment

Before that, we need to add variables to github repository configs.

Secrets:

• ACCESS_KEY
• CLUSTER_ID
• DOCKERHUB_REGISTRY
• DOCKERHUB_TOKEN
• DOCKERHUB_USERNAME
• PRIVATE_KEY

Variables:

• CONTAINER_NAME
• DEPLOYMENT_NAME
• IMAGE

10. Watch the status of the workflow. And verify the change by visit the service external ip.

11. Tear down the infrastructure after the demo

cd terraform
terraform destroy

Note

1. For the first time the user runs the terraform code, you might get a error which relate to AliyunCSManagedNlcRole. Follow the link to create this role so ACK can manage the pool. AliyunCSManagedNlcRole