Deploying Azure Data Factory

In my previous post I created an instance of Azure Data Factory to process files and backed it with Azure DevOps source control. In this post I’m deploying that Azure Data Factory to a production environment.

Code Repository

The important thing to remember from the previous post is that when an Azure Data Factory has a backing code repository setup, then each publish of the Azure Data Factory will push a set of ARM Template exports to the “adf_publish” branch automatically.

Git Data

adf_publish:

ADF Publish

New Data Factory

Setting up variables with names of things.

$currentTime = (Get-Date).ToUniversalTime().toString("yyyyMMdd-HHmm");
$armDeploymentName = $currentTime
$dataFactoryName = "adf-production";
$resourceGroupName = 'DataFactoryProd'

Creating a new resource group

$resourceGroup = New-AzResourceGroup $resourceGroupName -location 'East US'

Creating a new Azure Data Factory

$dataFactory = Set-AzDataFactoryV2 -ResourceGroupName $resourceGroup.ResourceGroupName -Location $resourceGroup.Location -Name $dataFactoryName

Now before deploying the pipelines and datasets and such, the ARM Parameter file in the repository must be filled out.

{
	"$schema": "https://schema.management.azure.com/schemas/2015-01-01/deploymentParameters.json#",
	"contentVersion": "1.0.0.0",
	"parameters": {
		"factoryName": {
			"value": "adf-production"
		},
		"DriveData_connectionString": {
			"value": "<Look in Azure Storage for this connection string>"
		},
		"SQL_Table_connectionString": {
			"value": "<Do you have a database connection string>"
		}
	}
}

Deploying files from the adf_publish branch of the repository, which I’ve cloned into a local directory and am running this following command from:

New-AzResourceGroupDeployment -Name $armDeploymentName -ResourceGroupName $resourceGroupName -TemplateFile .\ARMTemplateForFactory.json -TemplateParameterFile .\ARMTemplateParametersForFactory.json -Mode Incremental

Summary

Updating an Azure Data Factory instance from another environment is easy enough that automating job updates between environments is smooth. And with this powershell, it’s build server agnostic.