Skip to content

Latest commit

 

History

History
156 lines (127 loc) · 6.69 KB

multiple-instance-deploy.md

File metadata and controls

156 lines (127 loc) · 6.69 KB

Server Deployment Guide

This guide explains the GitHub Actions workflow for deploying to multiple servers and checking the deployment status using a health check endpoint. You can adapt this setup to deploy to multiple instances in different environments.

Disclaimer: Database Setup

Important: Before deploying your application, you need to set up a MySQL database separately. This guide does not cover the database setup process.

You can use Amazon RDS for MySQL or another provider for a managed database solution. After setting up your database, make sure to add the following environment variables to your instance:

  • MYSQL_HOST: The endpoint of your RDS instance
  • MYSQL_PORT: The port number (usually 3306 for MySQL)
  • MYSQL_DATABASE: The name of your database
  • MYSQL_USER: The username for accessing the database
  • MYSQL_PASSWORD: The password for the database user

These variables should be used in your application's configuration to connect to the database.

GitHub Actions Workflow

name: Deploy to Instances

on:
   push:
      branches:
         - main

jobs:
   deploy:
      runs-on: ubuntu-latest
      env:
         OCI_CLI_USER: ${{ secrets.OCI_CLI_USER }}
         OCI_CLI_TENANCY: ${{ secrets.OCI_CLI_TENANCY }}
         OCI_CLI_FINGERPRINT: ${{ secrets.OCI_CLI_FINGERPRINT }}
         OCI_CLI_KEY_CONTENT: ${{ secrets.OCI_CLI_KEY_CONTENT }}
         OCI_CLI_REGION: ${{ secrets.OCI_CLI_REGION }}
         API_URL: ${{ secrets.API_URL }}

      steps:
         - name: Restart Server 1 (Instance 1)
           uses: oracle-actions/run-oci-cli-command@v1.1.1
           id: restart_instance_1
           with:
              command: "compute instance action --instance-id ${{ secrets.OCI_INSTANCE_ONE_OCID }} --action SOFTRESET"

         - name: Allow time for server to restart
           run: sleep 300

         - name: Check Server 1 Health
           uses: oracle-actions/run-oci-cli-command@v1.1.1
           id: check_instance_1_health
           with:
              command: "lb backend-health get --backend-name ${{ secrets.OCI_INSTANCE_ONE_PRIVATE_IP_AND_PORT }} --backend-set-name ${{ secrets.OCI_LB_BACKEND_SET }} --load-balancer-id ${{ secrets.OCI_LB_OCID }}"

         - name: Verify Server 1 Health Status
           run: |
              OUTPUT='${{ steps.check_instance_1_health.outputs.output }}'

              # Parse the status from data object
              HEALTH_STATUS=$(echo "$OUTPUT" | jq -r 'fromjson | .data.status')
              echo "Health Status: $HEALTH_STATUS"
              if [ "$HEALTH_STATUS" != "OK" ]; then
                echo "Server 1 health check failed! Status: $HEALTH_STATUS"
                exit 1
              fi

         - name: Restart Server 2 (Instance 2)
           uses: oracle-actions/run-oci-cli-command@v1.1.1
           id: restart_instance_2
           with:
              command: "compute instance action --instance-id ${{ secrets.OCI_INSTANCE_TWO_OCID }} --action SOFTRESET"

         - name: Allow time for server to restart
           run: sleep 300

         - name: Check Server 2 Health
           uses: oracle-actions/run-oci-cli-command@v1.1.1
           id: check_instance_2_health
           with:
              command: "lb backend-health get --backend-name ${{ secrets.OCI_INSTANCE_TWO_PRIVATE_IP_AND_PORT }} --backend-set-name ${{ secrets.OCI_LB_BACKEND_SET }} --load-balancer-id ${{ secrets.OCI_LB_OCID }}"

         - name: Verify Server 2 Health Status
           run: |
              OUTPUT='${{ steps.check_instance_2_health.outputs.output }}'

              # Parse the status from data object
              HEALTH_STATUS=$(echo "$OUTPUT" | jq -r 'fromjson | .data.status')
              echo "Health Status: $HEALTH_STATUS"
              if [ "$HEALTH_STATUS" != "OK" ]; then
                echo "Server 2 health check failed! Status: $HEALTH_STATUS"
                exit 1
              fi

         - name: Check deployment
           run: |
              response=$(curl -s -o /dev/null -w "%{http_code}" ${API_URL}/api/health)
              if [ $response = "200" ]; then
                echo "Deployment successful!"
              else
                echo "Deployment failed!"
                exit 1
              fi

Workflow Explanation

  1. Trigger: The workflow runs when changes are pushed to the main branch.
  2. Deployment:
    • Restarts the first instance using the Oracle Cloud CLI
    • Waits for the server to restart
    • Checks the health status of the first instance
    • Restarts the second instance
    • Waits for the server to restart
    • Checks the health status of the second instance
    • Verifies the health status of both instances
    • Checks the deployment status by making a request to a health endpoint

General Setup Instructions

  1. Launch at least two instances:
    • Go to Oracle Cloud Console and create two instances.
    • Note down the OCID of each instance.
  2. Create a Load Balancer:
    • Create a load balancer in Oracle Cloud Console.
    • Add the instances to the backend set of the load balancer.
    • Note down the OCID of the load balancer.
  3. Set up Oracle Cloud CLI:
    • Configure the CLI with the required credentials.
  4. Set up GitHub Secrets:
    • In your GitHub repository settings, add these secrets:
      • OCI_CLI_USER: Oracle Cloud CLI user
      • OCI_CLI_TENANCY: Oracle Cloud tenancy OCID
      • OCI_CLI_FINGERPRINT: Fingerprint of the public key used for CLI
      • OCI_CLI_KEY_CONTENT: Private key content for CLI
      • OCI_CLI_REGION: Oracle Cloud region
      • OCI_INSTANCE_ONE_OCID: OCID of the first instance
      • OCI_INSTANCE_TWO_OCID: OCID of the second instance
      • OCI_INSTANCE_ONE_PRIVATE_IP_AND_PORT: Private IP and port of the first instance
      • OCI_INSTANCE_TWO_PRIVATE_IP_AND_PORT: Private IP and port of the second instance
      • OCI_LB_OCID: OCID of the load balancer
      • OCI_LB_BACKEND_SET: Name of the backend set in the load balancer
      • API_URL: The URL of your API. This is used to check if the deployment was successful
  5. Push to Main Branch or Run Manually:
    • Push your changes to the main branch to trigger the deployment workflow.

For further instructions on setting up the servers in oracle, you can refer to our setup markdown file here.

Best Practices

  • Secure your instance and keep GitHub secrets safe
  • Follow best practices for production deployments
  • Consider setting up staging environments
  • Implement comprehensive testing before production deployment

This setup enables continuous deployment, automatically updating your production environment with the latest code whenever you push to the master branch.