Daily Archives: April 22, 2019

Using Azure DevOps Build Pipeline Templates with Terraform to build an AKS cluster

Posted on by

In the last few posts I have moved from building an Azure SQL DB with Terraform using VS Code to automating the build process for the Azure SQL DB using Azure DevOps Build Pipelines to using Task Groups in Azure DevOps to reuse the same Build Process and build an Azure Linux SQL VM and Network Security Group. This evolution is fantastic but Task Groups can only be used in the same Azure DevOps repository. It would be brilliant if I could use Configuration as Code for the Azure Build Pipeline and store that in a separate source control repository which can be used from any Azure DevOps Project.

Luckily, you can 😉 You can use Azure DevOps Job Templates to achieve this. There is a limitation at present, you can only use them for Build Pipelines and not Release Pipelines.

The aim of this little blog series was to have a single Build Pipeline stored as code which I can use to build any infrastructure that I want with Terraform in Azure and be able to use it anywhere

Creating a Build Pipeline Template

I created a GitHub repository to hold my Build Templates, feel free to use them as a base for your own but please don’t try and use the repo for your own builds.

The easiest way to create a Build Template is to already have a Build Pipeline. This cannot be done from a Task Group but I still have the Build Pipeline from my automating the build process for the Azure SQL DB using Azure DevOps Build Pipelines blog post.

There is a View YAML button. I can click this to view the YAML definition of the Build Pipeline

I copy that and paste it into a new file in my BuildTemplates repository. (I have replaced my Azure Subscription information in the public repository)

jobs:
- job: Build
  pool:
    name: Hosted VS2017
    demands: azureps
  steps:
  - task: AzureCLI@1
    displayName: 'Azure CLI to deploy azure storage for backend'
    inputs:
      azureSubscription: 'PUTYOURAZURESUBNAMEHERE'
      scriptLocation: inlineScript
      inlineScript: |
        # the following script will create Azure resource group, Storage account and a Storage container which will be used to store terraform state
        call az group create --location $(location) --name $(TerraformStorageRG)
        
        call az storage account create --name $(TerraformStorageAccount) --resource-group $(TerraformStorageRG) --location $(location) --sku Standard_LRS
        
        call az storage container create --name terraform --account-name $(TerraformStorageAccount)

  - task: AzurePowerShell@3
    displayName: 'Azure PowerShell script to get the storage key'
    inputs:
      azureSubscription: 'PUTYOURAZURESUBNAMEHERE'
      ScriptType: InlineScript
      Inline: |
        # Using this script we will fetch storage key which is required in terraform file to authenticate backend stoarge account
      
        $key=(Get-AzureRmStorageAccountKey -ResourceGroupName $(TerraformStorageRG) -AccountName $(TerraformStorageAccount)).Value[0]
      
        Write-Host "##vso[task.setvariable variable=TerraformStorageKey]$key"
      azurePowerShellVersion: LatestVersion

  - task: qetza.replacetokens.replacetokens-task.replacetokens@3
    displayName: 'Replace tokens in terraform file'
    inputs:
      rootDirectory: Build
      targetFiles: |
        **/*.tf
        **/*.tfvars
      tokenPrefix: '__'
      tokenSuffix: '__'

  - powershell: |
      Get-ChildItem .\Build -Recurse
    
      Get-Content .\Build\*.tf 
      Get-Content .\Build\*.tfvars 
    
      Get-ChildItem Env: | select Name
    displayName: 'Check values in files'
    enabled: false

  - task: petergroenewegen.PeterGroenewegen-Xpirit-Vsts-Release-Terraform.Xpirit-Vsts-Release-Terraform.Terraform@2
    displayName: 'Initialise Terraform'
    inputs:
      TemplatePath: Build
      Arguments: 'init -backend-config="0-backend-config.tfvars"'
      InstallTerraform: true
      UseAzureSub: true
      ConnectedServiceNameARM: 'PUTYOURAZURESUBNAMEHERE'

  - task: petergroenewegen.PeterGroenewegen-Xpirit-Vsts-Release-Terraform.Xpirit-Vsts-Release-Terraform.Terraform@2
    displayName: 'Plan Terraform execution'
    inputs:
      TemplatePath: Build
      Arguments: plan
      InstallTerraform: true
      UseAzureSub: true
      ConnectedServiceNameARM: 'PUTYOURAZURESUBNAMEHERE'

  - task: petergroenewegen.PeterGroenewegen-Xpirit-Vsts-Release-Terraform.Xpirit-Vsts-Release-Terraform.Terraform@2
    displayName: 'Apply Terraform'
    inputs:
      TemplatePath: Build
      Arguments: 'apply -auto-approve'
      InstallTerraform: true
      UseAzureSub: true
      ConnectedServiceNameARM: 'PUTYOURAZURESUBNAMEHERE'

Now I can use this yaml as configuration as code for my Build Pipeline 🙂 It can be used from any Azure DevOps project. Once you start looking at the code and the documentation for the yaml schema you can begin to write your pipelines as YAML, but sometimes it is easier to just create build pipeline or even just a job step in the browser and click the view yaml button!

Create an AKS Cluster with a SQL 2019 container using Terraform and Build templates

I have a GitHub Repository with the Terraform code to build a simple AKS cluster. This could not have been achieved without Richard Cheney’s article I am not going to explain how it all works for this blog post or some of the negatives of doing it this way. Instead lets build an Azure DevOps Build Pipeline to build it with Terraform using Configuration as Code (the yaml file)

I am going to create a new Azure DevOps Build Pipeline and as in the previous posts connect it to the GitHub Repository holding the Terraform code.

This time I am going to choose the Configuration as code template

I am going to give it a name and it will show me that it needs the path to the yaml file containing the build definition in the current repository.

Clicking the 3 ellipses will pop-up a file chooser and I pick the build.yaml file

The build.yaml file looks like this. The name is the USER/Repository Name and the endpoint is the name of the endpoint for the GitHub service connection in Azure DevOps. The template value is the name of the build yaml file @ the name given for the repository value.

resources:
  repositories:
    - repository: templates
      type: github
      name: SQLDBAWithABeard/Presentations-BuildTemplates-Private
      endpoint: SQLDBAWithABeardGitHub

jobs:
- template: AzureTerraform.yaml@templates  # Template reference

You can find (and change) your GitHub service connection name by clicking on the cog bottom left in Azure DevOps and clicking service connections

I still need to create my variables for my Terraform template (perhaps I can now just leave those in my code?) For the AKS Cluster build right now I have to add presentation, location, ResourceGroupName, AgentPoolName, ServiceName, VMSize, agent_count

Then I click save and queue and the job starts running

If I want to edit the pipeline it looks a little different

The variables and triggers can be found under the 3 ellipses on the top right

It also defaults the trigger to automatic deployment.

It takes a bit longer to build

and when I get the Terraform code wrong and the build fails, I can just alter the code, commit it, push and a new build will start and the Terraform will work out what is built and what needs to be built!

but eventually the job finishes successfully

and the resources are built

and in Visual Studio Code with the Kubernetes extension installed I can connect to the cluster by clicking the 3 ellipses and Add Existing Cluster

I choose Azure Kubernetes Services and click next

Choose my subscription and then add the cluster

and then I can explore my cluster

I can also see the dashboard by right clicking on the cluster name and Open Dashboard

Right clicking on the service name and choosing describe

shows the external IP address, which I can put into Azure Data Studio and connect to my container

So I now I can source control my Build Job Steps and hold them in a central repository. I can make use of them in any project. This gives me much more control and saves me from repeating myself repeating myself. The disadvantage is that there is no handy warning when I change the underlying Build Repository that I will be affecting other Build Pipelines and there is no easy method to see which Build Pipelines are dependent on the build yaml file

Happy Automating

Using the same Azure DevOps build steps for Terraform with different Pipelines with Task Groups to build an Azure Linux SQL VM

Posted on by

In my last post I showed how to build an Azure DevOps Pipeline for a Terraform build of an Azure SQLDB. This will take the terraform code and build the required infrastructure.

The plan all along has been to enable me to build different environments depending on the requirement. Obviously I can repeat the steps from the last post for a new repository containing a Terraform code for a different environment but

If you are going to do something more than once Automate It

who first said this? Anyone know?

The build steps for building the Terraform are the same each time (if I keep a standard folder and naming structure) so it would be much more beneficial if I could keep them in a single place and any alterations to the process only need to be made in the one place 🙂

Task Groups

Azure DevOps has task groups. On the Microsoft Docs web-page they are described as


A task group allows you to encapsulate a sequence of tasks, already defined in a build or a release pipeline, into a single reusable task that can be added to a build or release pipeline, just like any other tas


https://docs.microsoft.com/en-us/azure/devops/pipelines/library/task-groups?view=azure-devops

If you are doing this with a more complicated existing build pipeline it is important that you read the Before You Create A Task Group on the docs page. This will save you time when trying to understand why variables are not available (Another grey hair on my beard!)

Creating A Task Group

Here’s the thing, creating a task group is so easy it should be the default way you create Azure DevOps Pipelines. Let me walk you through it

I will use the Build Pipeline from the previous post. Click edit from the build page

Then CTRL and click to select all of the steps

Right Click and theres a Create Task Group button to click !

You can see that it has helpfully added the values for the parameters it requires for the location, Storage Account and the Resource Group.

Remember the grey beard hair above? We need to change those values to use the variables that we will add to the Build Pipeline using

$(VariableName)

Once you have done that click Create

This will also alter the current Build Pipeline to use the Task Group. Now we have a Task Group that we can use in any build pipeline in this project.

Using the Task Group with a new Build Pipeline to build an Azure Linux SQL VM

Lets re-use the build steps to create an Azure SQL Linux VM. First I created a new GitHub Repository for my Terraform code. Using the docs I created the Terraform to create a resource group, a Linux SQL VM, a virtual network, a subnet, a NIC for the VM, a public IP for the VM, a netwwork security group with two rules, one for SQL and one for SSH. It will look like this

The next step is to choose the repository

again we are going to select Empty job (although the next post will be about the Configuration as Code 🙂

As before we will name the Build Pipeline and the Agent Job Step and click the + to add a new task. This time we will search for the Task Group name that we created

I need to add in the variables from the variable.tf in the code and also for the Task Group

and when I click save and queue

It runs for less than 7 minutes

and when I look in the Azure portal

and I can connect in Azure Data Studio

Altering The Task Group

You can find the Task Groups under Pipelines in your Azure DevOps project

Click on the Task Group that you have created and then you can alter, edit it if required and click save

This will warn you that any changes will affect all pipelines and task groups that are using this task group. To find out what will be affected click on references


which will show you what will be affected.

Now I can run the same build steps for any Build Pipeline and alter them all in a single place using Task Groups simplifying the administration of the Build Pipelines.

The next post will show how to use Azure DevOps templates to use the same build steps across many projects and build pipelines and will build a simple AKS cluster

The first post showed how to build an Azure SQLDB with Terraform using VS Code

The second post showed how to use Azure DevOps Task Groups to use the same build steps in multiple pipelines and build an Azure Linux SQL Server VM

Happy Automating!