Simple Demo for AWS CICD Pipeline

In this article we will use a simple C# code to present how CICD pipeline build by using AWS CodeBuild, CodeDeploy and CodePipeline. Fortunately, in last article, we have already build the CI pipeline using CodeBuild with other services, therefore, now we can use the pipeline and continue build the left part.

0. Simple Code Prepare (DONE)

1. Build Stage (CHANGE)

If you remember what we have done in last article, we use the AWS CodeBuild panel and select the output destination as S3 bucket and the type is zip file, this is one way to define a build procedure, we can also define it inside the buildspec.yml.

1.1 change the artifacts in buildspec.yml file

Here we need to notice that, because the CodeDeploy step, we need the appspec.yml file and also some scripts, so at artifacts, we need to not only include the output/ directry, but also with these necessary files.

artifacts:
  files:
    - appspec.yml
    - scripts/*
    - output/**/*
  discard-paths: no  # Keep folder structure

1.2 appspec.yml file

appsepc.yml is used to tell codedeploy-agent what to do with the files, later we will see how to install codedeploy-agent on the EC2 instance.

In the GitHub repo, add appspec.yml for CodeDeploy:

version: 0.0
os: linux
files:
  - source: /
    destination: /home/ubuntu/DemoConsoleApp
permissions:
  - object: /home/ubuntu/DemoConsoleApp/scripts
    pattern: "*.sh"
    owner: ubuntu
    mode: 755
hooks:
  BeforeInstall:
    - location: scripts/before_install.sh
      timeout: 300
      runas: ubuntu
  ApplicationStart:
    - location: scripts/start.sh
      timeout: 300
      runas: ubuntu

Here mainly we do 3 things:

• Deploy/copy all files to /home/ubuntu/DemoConsoleApp on EC2.
• Sets correct permissions (chmod +x) for the *.sh file.
• Runs start.sh automatically after deployment.

Notice, here we don't need to manually copy the appspec.yml or scripts to the folder where we need (eg. /home/ubuntu/DemoConsoleApp/). This is what codedeploy-agent have done according to appspec.yml file.

1.3 Create start.sh for Running the App

Inside our GitHub repo, we also need to add a scripts/start.sh file:

#!/bin/bash
cd /home/ubuntu/DemoConsoleApp/output
nohup dotnet DemoConsoleApp.dll > app.log 2>&1 &
echo "App started successfully!"

and also a before_install.sh script to garantee the folder exists.

#!/bin/bash

set -e

TARGET_DIR="/home/ubuntu/DemoConsoleApp"

echo "Checking and creating directory: $TARGET_DIR"

if [ ! -d "$TARGET_DIR" ]; then
    mkdir -p "$TARGET_DIR"
    echo "Directory created successfully."
elif [ -d "$TARGET_DIR" ]; then
    echo "Directory already exists."
else
    echo "Error: $TARGET_DIR exists but is not a directory."
    exit 1
fi

2. Deploy Stage

2.1 Install codedeploy-agent on EC2 instance

login to EC2 instance and install codedeploy-agent on the instance.

# ubuntu
sudo apt update -y
sudo apt install ruby-full -y
sudo apt install wget -y
cd /home/ubuntu
wget https://aws-codedeploy-eu-central-1.s3.eu-central-1.amazonaws.com/latest/install
chmod +x ./install
sudo ./install auto
sudo systemctl start codedeploy-agent
sudo systemctl enable codedeploy-agent
sudo systemctl status codedeploy-agent

here as I used ubuntu instance, I used above commands. If you use Redhat-like instance, you can use yum cmd, more info about codedeploy-agent pls check here).

Now codedeploy-agent is running

2.2 Create role for EC2 (if not) and CodeDeploy

Attach the AmazonEC2RoleforAWSCodeDeploy and AWSCodeDeployFullAccess policy to the EC2 instance role.

Also for CodeDeploy, we need another role, with policy AWSCodeDeployRole.

2.3 Create an Application in CodeDeploy

2.3.1 Create an Application

• Application Name: DemoConsoleAppDeploy

2.3.2 Create the Deployment Group

• Deployment Group Name: DemoConsoleApp-Group
• Environment configuration:
  • Amazon EC2 instances
• Service Role: Attach the role we created above
• Deployment Type: In-Place
  • In-Place will stop the app on EC2, updates it, and restarts it.
  • Blue/Green will create new instances and routes traffic to them.
• Deployment Strategy: All at Once
  • One at a Time: there will be no downtime, but it will be slow
  • Half at a Time
  • All at Once: Fastest, but downtime occurs, as here, this is not so critical as a demo example.
• Load Balancer: Disabled

2.3.3 Create the Deployment

• Revision:
  • A revision is the application package that CodeDeploy fetches and deploys to our EC2 instance. Here we can use the previous s3 bucket.

3. CodePipeline

As we already have setup the source, and connected source provider Github, and also setup the CodeBuild project, then as above show, we have setup the CodeDeploy application. Therefore in the CodePipeline, we just need to choose each stages provider and choose corresponding items.

• Pipeline Name: demo-consoleapp-pipeline
• Source Provider: GitHub
• Build Provider: CodeBuild
• Deploy Provider: CodeDeploy

Then we can start the pipeline. And after some time, we can see the deploy is successful finally.

4. Error Analysis

4.1 Build Stage Errors

For Build Stage, the error we can directly checked in the panel, "Build logs" section.

4.2 Deploy Stage Errors

Here as we deployed on EC2 instance, so we can check the corresponding logs on the instance. The default codedeploy-agent logs stored at /var/log/aws/codedeploy-agent/ dir.

tail -100 /var/log/aws/codedeploy-agent/codedeploy-agent.log

And the standard output/error log we can also check here:

cat /opt/codedeploy-agent/deployment-root/deployment-logs/codedeploy-agent-deployments.log

And we can also find the file stored place where it copyed from s3 bucket /opt/codedeploy-agent/deployment-root.

find /opt/codedeploy-agent/deployment-root -maxdepth 3 -type d -name deployment-archive

Summary

Here the whole workflow is like: