A full devops project to deploy my Asp.net core web app (news app) with mysql database in a full gitflow including 5 branches (test, dev, preprod, release) and the main branch to contain the base code. Using these technologies and tools:

• GCP (Google Cloud Platform)
• Dotnet core 7
• Mysql
• Terraform
• Ansible
• Helm
• Jenkins
• Docker
• Kubernates
• Github

You can find all the detailed steps on the readme files on the following repos 👇🏽

Infrastructure: https://github.com/mohamedsamirspot/Full-Devops-Project_Infrastructure

Dotnet App: https://github.com/mohamedsamirspot/Full-Devops-Project_Dotnetcoreapp

Full-Devops-Project_Infrastructure:

in this infrastructure we are going to prepair the gcp infra using terraform for our kubernates cluster to deploy our jenkins and our app to it.

Demo.mp4

GCP Infrastructure includes:

• 1 VPC
• Backend for saving the tfstate file of terraform
• 2 subnets (management subnet & restricted subnet):
• NAT gateway over the restricted subnet
• Management subnet has the following:
  • Public VM
• Restricted subnet has the following:
  • private standard GKE cluster (private control plane and private worker nodes)

Firstly creating the required service account for our terraform

• Creating my key to access the public vm (bastian host) using ssh for ansible

  ssh-keygen -t ed25519 -C  <remote-user-name> -f /Terraform-folderpath/full-devops-proj-bastian-host-key/<file-name>
  

now copy the contents of the public key

• Now put the public key content in the public vm creation key part

• Now run the script.sh bash file to create the infra using terraform (including the bucket backend to store the state file) and configure the bastian host also using ansible and deploying the jenkins infra using ansible and helm

        ./script.sh
  

in case of you saw this error

empty this file (known_hosts)

• Now remove your local terraform.tfstate and terraform.tfstate.backup files you don't need them

• Build and push the jenkins slave dockerfile with all the required tools installed on it --> Salve-Pod-Dockerfile

  docker build --build-arg JENKINS_PASSWORD=<myubuntu-jenkins-userpassword> -t mohamedsamirebrahim/enhanced-slave-image:latest -f Salve-Pod-Dockerfile .
  docker push mohamedsamirebrahim/enhanced-slave-image:latest
  

• Now ssh to this public vm to get the public ip of the jenkins service load balancer

      kubectl get all -n jenkins
  

• And now at this point we have our infrastructure up and running with jenkins and its slaves as deployment on our GKE cluster
• Now access master jenkins pod using the service loadbalancer ip

  kubectl exec -it jenkins-master-pod-name -n jenkins -- bash
  cat /var/jenkins_home/secrets/initialAdminPassword

or just get it from logs

  kubectl logs jenkins-master-dep-6c68d86f64-qqvq2 -n jenkins

• preparing the branches of your app including the jenkins file, docker file and the helm chart to deploy the app

• Go to security and enable proxy compatibility

• Now go to credentials and create username and password credential for the slave jenkins user that we created before in the dockerfile of the slave

• Now set up the jenkins agent (slave)

• Create dockerhub credential

• Get the kubeconfig from any of the connectors to the api-server (node or pod that runs this before gcloud container clusters get-credentials <privatecluster_name> --zone <zone_name> --project <project_id>)

    cat ~/.kube/config
  

• now create a kubeconfig file with this content

• then create a secretfile credential to use this kubeconfig file content on Jenkinsfile as the service account is working inside the pod w can create anything but it doesn't work inside the jenkins file so it is useless in our project 😞 it require a plugin

• create the database credential for mysql of type secret text (❗❗Note: must be in base 64 format)
• which will be passed in the jenkinsfile helm command as a value variable to be passed to the creation of the db-secret yaml file in the cluster, then to be passed to the app, database deployments containers
• in my case i will create 4 secrets one in each namespace and use it through the current env

create multibranch pipeline