I wrote an article on converting my Cosmos DB PowerShell module to use a CI process in Azure DevOps over on the awesome PowerShell Magazine. Check it out here.
Recently, I’ve been putting together a continuous delivery pipeline (using VSTS) for our Azure API Management service using Azure Resource Manager (ARM) templates. One of the things I needed to be able to do to secure this service properly is to disable TLS 1.0, TLS 1.1 and 3DES. This is pretty easy to do in the portal:
However, we only allow changes to be made via our continuous delivery pipeline (a good thing by the way) then I had to change the ARM template.
Side note: Disabling TLS 1.0, TLS 1.1 and 3DES is pretty important for keeping your system secure. But if you have an Azure Application Gateway in front of your API Management service, then you’ll also need to configure the Azure Application Gateway to disable TLS 1.0 and TLS 1.1. This is done in a slightly different way, but can also be done in an ARM Template (post a comment if you’re not sure how to do this and I’ll write another post).
I found the documentation for the API Management service resource here. This shows it can be done by setting the customProperties object in the ARM Template. But the documentation isn’t completely clear.
But after a little bit of trial and error I managed to figure it out and get it working. What you need to do is add the following customProperties to the properties of the API Management service resource:
This is what the complete ARM template looks like:
Side note: the template above is based off the Azure Quickstart Template for API Management.
Hopefully you find this if you’re looking for an example of how to do this and it saves you some time.
If you’re just getting started with Cosmos DB, you might not have come across users and permissions in a Cosmos DB database. However, there are certain use cases where managing users and permissions are necessary. For example, if you’re wanting to be able to limit access to a particular resource (e.g. a collection, document, stored procedure) by user.
The most common usage scenario for users and permissions is if you’re implementing a Resource Token Broker type pattern, allowing client applications to directly access the Cosmos DB database.
Side note: The Cosmos DB implementation of users and permissions only provides authorization – it does not provide authentication. It would be up to your own implementation to manage the authentication. In most cases you’d use something like Azure Active Directory to provide an authentication layer.
But if you go hunting through the Azure Management Portal Cosmos DB data explorer (or Azure Storage Explorer) you won’t find any way to configure or even view users and permissions.
But to make Cosmos DB users and permissions easier to manage from PowerShell, I created the Cosmos DB PowerShell module. This is an open source project hosted on GitHub. The Cosmos DB module allows you to manage much more than just users and permissions, but for this post I just wanted to start with these.
This module works on PowerShell 5.x and PowerShell Core 6.0.0. It probably works on PowerShell 3 and 4, but I don’t have any more machines running this version to test on.
The Cosmos DB module does not have any dependencies, except if you call the New-Cosmos DbContext function with the ResourceGroup parameter specified as this will use the AzureRM PowerShell modules to read the Master Key for the connection directly from your Cosmos DB account. So I’d recommend installing the Azure PowerShell modules or if you’re using PowerShell 6.0, install the AzureRM.NetCore modules.
Installing the Module
Install-Module -Name CosmosDB -Scope CurrentUser
Or to install it for all users on the machine (requires administrator permissions):
Install-Module -Name CosmosDB
As of Cosmos DB module v2.0.1, the connection parameter has been renamed to context and the New-CosmosDbConnection function has been renamed New-CosmosDbContext. This was to be more inline with naming adopted by the Azure PowerShell project. The old connection parameters and New-CosmosDbConnection function is still available as an alias, so older scripts won’t break. But these should be changed to use the new naming if possible as I plan to deprecate the connection version at some point in the future.
This post was updated to specify the new naming, but screenshots still show the Connection aliases.
Before you get down to the process of working with Cosmos DB resources, you’ll need to create a context variable containing the information required to connect. This requires the following information:
- The Cosmos DB Account name
- The Cosmos DB Database name
- The Master Key for the account (you can have the Cosmos DB PowerShell module get this directly from your Azure account if you wish).
To create the connection variable we just use the New-CosmosDbContext:
If you do not wish to specify your master key, you can have the New-CosmosDbContext function pull your master key from the Azure Management Portal directly:
Note: This requires the AzureRM.Profile and AzureRM.Resoures module on Windows PowerShell 5.x or AzureRM.Profile.NetCore and AzureRM.Resources.NetCore on PoweShell Core 6.0.0.
To add a user to the Cosmos DB Database use the New-CosmosDBUser function:
New-CosmosDbUser -Context $context -Id 'daniel'
To get a list of users in the database:
Get-CosmosDbUser -Context $context
To get a specific user:
Get-CosmosDbUser -Context $context -Id 'daniel'
To remove a user (this will also remove all permissions assigned to the user):
Remove-CosmosDbUser -Context $context -Id 'daniel'
Permissions in Cosmos DB are granted to a user for a specific resource. For example, you could grant a user access to just a single document, an entire collection or to a stored procedure.
To grant a permission you need to provide four pieces of information:
- The Id of the user to grant the permission to.
- An Id for the permission to create. This is just string to uniquely identify the permission.
- The permission mode to the permission: All or Read.
- The Id of the resource to grant access to. This can be generated from one of the Get-CosmosDb*ResourcePath functions in the CosmosDB PowerShell module.
In the following example, we’ll grant the user daniel all access to the TestCollection:
Once a permission has been granted, you can use the Get-CosmosDbPermission function to retrieve the permission and with it the Resource Token that can be used to access the resource for a limited amount of time (between 10 minutes and 5 hours).
Note: as you have the Master Key already, using the Resource Token isn’t required.
For example, to retrieve all permissions for the user with Id daniel and a resource token expiration of 600 seconds:
Get-CosmosDbPermission -Context $context -UserId 'daniel' -TokenExpiry '600' | fl *
You can as expected delete a permission by using the Remove-CosmosDbPermission function:
Remove-CosmosDbPermission -Context $context -UserId 'daniel' -Id 'AccessTestCollection'
So this is pretty much all there is to managing users and permissions using the Cosmos DB PowerShell module. This module can also be used to manage the following Cosmos DB resources:
- Stored procedures
- User Defined Functions
You can find additional documentation and examples of how to manage these resources over in the Cosmos DB PowerShell module readme file on GitHub.
Hopefully this will help you in any Cosmos DB automation tasks you might need to implement.
Here is a short (and sometimes handy) single line of PowerShell code that can be used to restart all the Azure Web Apps in a subscription:
Note: Use this with care if you’re working with production systems because this _will_ restart these Web Apps without confirming first.
This would be a handy snippet to be able to run in the Azure Cloud Shell. It could also be adjusted to perform different actions on other types of resources.
To stop all Web Apps in a subscription use:
To start them all again:
The key part of this command is the GetEnumerator() method because most Azure Cmdlets don’t return an array of individual objects into the pipeline like typical PowerShell cmdlets. Instead returning a System.Collections.Generic.List object, which requires a slight adjustment to the code. This procedure can be used for most Azure Cmdlets to allow the results to be iterated through.
Thanks for reading.
The Azure PowerShell cmdlets are really first class if you’re wanting to manage Azure with PowerShell. However, they don’t always support the very latest Azure components and features. For example, at the time of writing this there is no Azure PowerShell module for managing Azure Container Instances.
The solution to this is to install the Nightly Build of Azure CLI 2.0. However, on Windows it is not entirely clear the easiest way to do this. So, in this post I’ll provide a PowerShell script that will:
- Install Python 3.x using Chocolatey
- Use PIP (Python package manager) to install the latest nightly build packages
- Update the Environment Path variable so that you can use Azure CLI 2.0.
Note: If you have the stable build of Azure CLI 2.0 installed using the MSI then you’ll need to configure your Environment Path variable to find the Az command that you’d like to use by default. I personally removed the stable build of Azure CLI 2.0 to make it easier.
Performing the Install
Make sure you’ve got Chocolatey installed. If you aren’t sure what Chocolatey is, it is a package management system for Windows – not unlike Apt-Get or Yum for Linux. It is free and awesome. In this process we’ll use Chocolatey to install Python for us. If you haven’t got Chocolatey installed, see this page for instructions.
Next, download and run this PowerShell script in a PowerShell Administrator Console:
You could save the content of this script into a PS1 file and then execute it like this:
It will then download and install Python, then use PIP to install the current nightly build packages. After a few minutes the installation will complete:
You can then run:
To get started.
If you’re a bit new to Azure CLI 2.0, then another great way is to use Azure CLI Interactive:
If you need to update to a newer nightly build, just run the script again and it will update your packages.
Easy as that! Now you can experiment with all the latest automation features in Azure without needing to wait for a new version of Azure CLI 2.0 or for latest Azure PowerShell cmdlets.
If you want to install even more “bleeding edge” builds (built straight off the master branch on every merge to master) then you can make a small adjustment to the script above:
On line 34 change the URL of the feed from:
Thanks for reading!
Note: I don’t actually plan on running any of these apps, I just wanted to test out the process and my scripts to identify any problems.
- Sonatype Nexus 2 – sonatype/nexus:oss
- Sonatype Nexus 3 – sonatype/nexus3:latest
- Jenkins – jenkins/jenkins
And here are the results of my tests:
Sonatype Nexus 2
Works perfectly and the container starts up quickly (under 10 seconds):
I passed the following parameters to the script:
Note: The Nexus 2 server is only accessible on the path /nexus/.
Sonatype Nexus 3
Works perfectly but after takes at least a minute to be accessible after the container starts. But this is normal behavior for Nexus 3.
I passed the following parameters to the script:
Unfortunately Jenkins does not work with a persistent storage volume from an Azure Share. It seems to be trying to set the timestamp of the file that will contain the InitialAdminPassword, which is failing:
I passed the following parameters to the script:
So, this is still a little bit hit and miss, but in general Azure Container Instances look like a very promising way to run different types of services in containers without a lot of overhead. With a bit of automation, this could turn out to be a cost effective way to quickly and easily run some common services.
Update 2018-04-26: At some point Microsoft made a change to the requirements of the ARM template creating the Azure Container Instance. It now requires the Ports to be specified within the container as well as we the container group. I have improved the ARM template to meet the current requirements.
Update 2017-08-06: I have improved the script so that it is idempotent (can be run more than once and will only create anything that is missing). The Azure Container Instance resource group can be deleted once you’ve finished with the container and then recreated again with this same script when you next need it. The storage will be preserved in the separate storage account resource group. The script can now be run with the -verbose parameter and will produce much better progress information.
Azure Container Instances (ACI) is a new resource type in Azure that allows you to quickly and easily create containers without the complexity or overhead of Azure Service Fabric, Azure Container Services or provisioning a Windows Server 2016 VM.
It allows you to quickly create containers that are billed by the second from container images stored in Docker Hub or your own Azure Container Registry (ACR). Even though this feature is still in preview, it is very easy to get up and running with it.
But this post isn’t about creating basic container instances, it is about running container instances where some of the storage must persist. This is a basic function of a container host, but if you don’t have access to the host storage then things get more difficult. That said, Azure Container Instances do support mounting Azure File Shares into the container as volumes. It is fairly easy to do, but requires quite a number of steps.
There is some provided documentation for persisting storage in a container instance, but it is quite a manual process and the example ARM templates are currently broken: there are some typos and missing properties. So this post aims to make the whole thing a lot simpler and automatable.
So in this post, I’m going to share a PowerShell function and Azure Resource Manager (ARM) template that will allow you to easily provision an Azure Container Instance with an Azure File Share mounted. The process defaults to installing a GoCD Server container (version 17.8.0 if you’re interested), but you could use it to install any other Linux Container that needs persistent storage. The script is parameterized so other containers and mount points can be specified – e.g. it should be fairly easy to use this for other servers like Sonatype Nexus or Jenkins Server.
Update 2017-08-06: I documented my findings trying out these other servers in my following blog post.
To perform this process you will need the following:
- PowerShell 5.0+ (PowerShell 4.0 may work, but I haven’t tested it).
- The Azure PowerShell module installed.
- Created an Application Service Principal – see below.
Azure Service Principal
Before you start this process you will need to have created an Application Service Principal in Azure that will be used to perform the deployment. Follow the instructions on this page to create an application and then get the Service Principal from it.
You will need to record these values as they will be provided to the script later on:
- Application Id
- Application Key
- Tenant Id
- Subscription Name
The process will perform the following tasks:
- The Service Principal is used to login to Azure to perform the deployment.
- An Azure Resource Group is created to contain a Azure Storage Account and Azure Key Vault.
- An Azure Storage Account is created and an Azure File Share is created in it.
- An Azure Key Vault is created to store the Storage Account Key and make it accessible to the Azure Container Instance.
- The Service Principal is granted permission to the Azure Key Vault to read and write secrets.
- The key to the Storage Account Key is added as a secret to the Azure Key Vault.
- The parameters are set in an ARM Template parameter file.
- An Azure Resource Group is created to contain the Azure Container Instance.
This is the content of the script:
The script requires a four parameters to be provided:
- ServicePrincipalUsername – the Application Id obtained when creating the Service Principal.
- ServicePrincipalPassword – the Application Key we got (or set) when creating the Service Principal.
- TenancyId – The Tenancy Id we got during the Service Principal creation process.
- SubscriptionName – the name of the subscription to install the ACI and other resources into.
There are also some other optional parameters that can be provided that allow the container image that is used, the TCP port the container listens on and mount point for the Auzre File Share. If you don’t provide these parameters will be used which will create a GoCD Server.
- AppCode – A short code to identify this application. It gets added to the resource names and resource group names. Defaults to ‘gocd’.
- UniqueCode – this string is just used to ensure that globally unique names for the resources can be created. Defaults to ‘zzz‘.
- ContainerImage – this is the name and version of the container image to be deployed to the ACI. Defaults to ‘gocd/gocd-server:v17.8.0‘.
- CPU – The number of cores to assign to the container instance. Defaults to 1.
- MemoryInGB – The amount of memory (in GB) to assign to the container instance. Defaults to 1.5.
- ContainerPort – The port that the container listens on. Go CD Server defaults to 8153.
- VolumeName – this is a volume name that is used to represent the volume in the ARM template. It can really be set to anything. Defaults to ‘gocd‘.
- MountPoint – this is the folder in the Container that the Azure File Share is mounted to. Defaults to ‘/godata/‘.
ARM Template Files
There are two other files that are required for this process:
- ARM template – the ARM template file that will be used to install the ACI.
- ARM template parameters – this file will be used to pass in the settings to the ARM Template.
This file is called aci-azuredeploy.json and should be downloaded to the same folder as the script above.
ARM Template Parameters
This file is called aci-azuredeploy.parameters.json and should be downloaded to the same folder as the script above.
To use the script the following steps need to be followed:
- Download the three files above (the script and the two ARM template files) and put them into the same folder:
- Open a PowerShell window.
- Change directory to the folder you place the files into by executing:
- CD <folder location>
- Execute the script like this (passing in the variables):
- The process will then begin and make take a few minutes to complete:Note: I’ve changed the keys to this service principal and deleted this storage account, so I using these Service Principal or Storage Account keys won’t work!
- Once completed you will be able to log in to the Azure Portal and find the newly created Resource Groups:
- Open the resource group *gocdacirg and then select the container group *gocdaci:
- The IP Address of the container is displayed. You can copy this and paste it into a browser window along with the port the container exposed. In the case of Go CD it is 8153:
- The process is now completed.
The Azure Container Instance can now be deleted and recreated at will, to reduce cost or simply upgrade to a new version. The Azure File Share will persist the data stored by the container into the mounted volume:
Hopefully this process will help you implement persisted storage containers in Azure Container Instances more easily and quickly.
Thanks for reading!