In my last blog post I showed how to trigger Azure Data Factory (ADF) pipelines from Azure Automation. I also mentioned the option to process an Azure Analysis Services cube from Azure Automation. For example right after your ADF data processing finishes, which will probably be a common use case. In this blog post I show you how you can use the Analysis Services PowerShell provider, also known as SQLASCDMLETS, from Azure Automation.
Create custom SQLASCMDLETS module
The SQLASCDMLETS are not (yet) available in the PowerShell Gallery so unfortunately it’s not possible to import the cmdlets straight into Automation like I did with the ADF cmdlets in my previous blog post. Instead we have to create our own module which will contain the SQLASCMDLETS and its dependencies.
UPDATE: Most people find it easier to use a Logic App to process Azure Analysis Services objects. I wrote a new blog post that describes this approach, including scaling up before processing and scaling down afterwards. You can find the blog post here: https://jorgklein.com/2018/02/28/process-azure-analysis-services-objects-using-a-logic-app-part-2/
Alternatively, you can now use the SqlServer PowerShell Gallery module. Creating a custom module like described below in the blue text below is not required anymore. Instead, follow the 3 steps below and continue reading below the blue text.
1. Navigate to https://www.powershellgallery.com/packages/SqlServer/
2. Click the “Deploy to Azure Automation” button
3. Import the SqlServer module to your Azure Automation account
The required files come with SQL Server Management Studio (SSMS) which you can download and install for free. It’s important to note you need the latest version (140) of the SQLASCDMLETS which is shipped with the latest Release Candidate of SSMS. Download and install it.
If you try to use the previous version of the SQLASCDMLETS (130) you will get an error in Automation because it tries to authenticate with a claims token while only windows authentication is supported by the 130 version of SQLASCDMLETS: “The value ‘ClaimsToken’ is not supported for the connection string property”.
After installing SSMS you should now be able to see the following directory: C:\Program Files (x86)\Microsoft SQL Server\140\Tools\PowerShell\Modules\SQLASCMDLETS
Copy the SQLASCMDLETS folder to a temporary location, for example C:\SQLASCMDLETS.
You will need the following files:
We also need the libraries SQLASCMDLETS depends on. Search your computer for the following files and copy paste them to the C:\SQLASCMDLETS folder. Make sure you copy them from a folder that has “140” in the path so you are sure you have the correct version.
Now zip the entire folder, make sure the name is “SQLASCMDLETS.zip”.
Import custom SQLASCMDLETS module to Azure Automation
Navigate to your Azure Automation account in the Azure portal.
Click Add a module to import a custom module:
Now upload the SQLASCMDLETS.zip file:
The zip file will be extracted:
Wait until the extraction finished and the status changes to Available. Click on the module name:
You now see the available activities including the one we will use to process the Azure Analysis Services Database.
Create Azure Automation Credential
Now we need to create a Credential to be able to automatically login and run our PowerShell script unattended from Azure Automation.
Click “Add a credential” and enter an organization account that has permissions to process your Azure Analysis Services database. Make sure you enter the User Principal Name (UPN) and not a Windows AD account. It is often the email address and may look like firstname.lastname@example.org. Give the new Credential a name, I chose “adpo-auto-cred”. It will be referenced in the PowerShell script below.
Create Automation Runbook
You can use the simple PowerShell script below to process your Azure Analysis Services database from Azure Automation. It will use the “adpo-auto-cred” credential to authenticate and will process your database using the Invoke-ProcessASDatabase SQLASCMDLETS function.
Replace “dbname” with your database name and “server” with your server, e.g. asazure://westeurope.asazure.windows.net/yourserver and you are good to go.
Copy/paste the script below to a Windows PowerShell Script (.ps1) file and name it “ProcessASDatabase.ps1”.
$AzureCred = Get-AutomationPSCredential -Name "adpo-auto-cred" Add-AzureRmAccount -Credential $AzureCred | Out-Null Invoke-ProcessASDatabase -databasename "dbname" -server "server" -RefreshType "Full" -Credential $AzureCred
That’s it, you now have an Azure Automation Runbook that you can schedule, monitor and integrate with your other data platform related tasks!