Deploy Rancher on Azure for Kubernetes Management

Lately I have been hearing a lot about a solution named Rancher in the Kubernetes space. Rancher is an open source Kubernetes Multi-Cluster Operations and Workload Management solution. You can learn more about Rancher here: https://www.rancher.com.

In short you can use Rancher to deploy and manage Kubernetes clusters deployed to Azure, AWS, GCP their managed Kubernetes offerings like GCE, EKS, AKS or even if you rolled your own. Rancher also integrates with a bunch of 3rd party solutions for things like authentication such as Active Directory, Azure Active Directory, Github, and Ping and logging solutions such as Splunk, Elasticsearch, or a Syslog endpoint.

Recently training opened up for some Rancher/Kubernetes/Docker training so I decided to go. The primary focus was on Rancher while also covering some good info on Docker and Kubernetes. This was really good training with a lot of hands on time, however there was one problem with the labs. The labs had instructions and setup scripts ready to go to run Rancher local on your laptop or on AWS via Terraform. There was nothing for Azure.

I ended up getting my Rancher environment running on Azure but it would have been nice to have some scripts or templates ready to go to spin up Rancher on Azure. I did find some ARM templates to spin up Rancher but they deployed an old version and it was not clear in the templates on where they could be updated to deploy the new version of Rancher. I decided to spend some time building out a couple of ARM templates that can be used to quickly deploy Rancher on Azure and add a Kubernetes host to Rancher. In the ARM template I pulled together it pulls the Rancher container from Docker Hub so it will always deploy the latest version. In this blog post I will spell out the steps to get your Rancher up and running in under 15 minutes.

First off you can find the ARM Templates here on my Github here: https://github.com/Buchatech/DeployRanchertoAzure.

The repository consists of ARM templates for deploying Rancher and a host VM for Kubernetes. NOTE: These templates are intended for labs to learn Rancher. They are not intended for use in production.

In the repo ARM Template #1 named RancherNode.JSON will deploy an Ubuntu VM with Docker and the latest version of Rancher (https://hub.docker.com/r/rancher/rancher) from Docker Hub. ARM Template #2 named RancherHost.JSON will deploy an Ubuntu VM with Docker to be used as a Kubernetes host in Rancher.

Node Deployment

Deploy the RancherNode.JSON ARM template to your Azure subscription through “Template Deployment” or other deployment method. You will be prompted for the following info shown in the screenshot:

Host Deployment

Deploy the RancherHost.JSON ARM template to your Azure subscription through “Template Deployment” or other deployment method. Note that that should deploy this into the same Resource Group that you deployed the Rancher Node ARM template into. You will be prompted for the following info shown in the screenshot:

After the Rancher Node and Rancher Host ARM templates are deployed you should see the following resources in the new Resource Group:

NameType
RancherVNet Virtual network
RancherHost Virtual machine
RancherNode Virtual machine
RancherHostPublicIP Public IP address
RancherNodePublicIP Public IP address
RancherHostNic Network interface
RancherNodeNic Network interface
RancherHost_OSDisk Disk
RancherNode_OSDisk Disk

Next navigate the Rancher portal in the web browser. The URL is the DNS name of the Rancher Node VM. You can find the DNS name by clicking on the Rancher Node VM in the Azure portal on the overview page. Here is an example of the URL:

https://ranchernode.centralus.cloudapp.azure.com

The Rancher portal will prompt you to set a password. This is shown in the following screenshot.

After setting the password the Rancher portal will prompt you for the correct Rancher Server URL. This will automatically be the Rancher Node VM DNS name. Click Save URL.

You will then be logged into the Rancher portal. You will see the cluster page. From here you will want to add a cluster. Doing this is how you add a new Kubernetes cluster to Rancher. In this post I will show you how to add a cluster to the Rancher Host VM. When it’s all said and done Rancher will have successfully deployed Kubernetes to the Rancher Host VM. Note that you could add a managed Kubernetes such as AKS but we won’t do that in this blog. I will save that for a future blog post!

Click on Add Cluster

Under “From my own existing nodes” Click on custom, give the cluster a name and click Next.

Next check all the boxes for the Node Options since all the roles will be on a single Kubernetes cluster. Copy the code shown at the bottom of the page, click done and run the code on the Rancher Host.

In order to run the code on the Rancher Host you need to SSH in and run it from there. To do this follow these steps:

  1. In the Azure Portal, from within the resource group click on the Rancher Host VM.
  2. On the Overview page click on Connect.
  3. Copy “ssh ranchuser@rancherhost.centralus.cloudapp.azure.com” from the Connect to virtual machine pop up screen.
  4. Open a terminal in either Azure cloud shell or with something like a terminal via VS Code and past the “ssh ranchuser@rancherhost.centralus.cloudapp.azure.com” in.

Running the code will look like this:

When done you can run Docker PS to see that the Rancher agent containers are running.

In the Rancher portal under clusters you will see the Rancher host being provisioned

The status will change as Kubernetes is deployed.

Once it’s done provisioning you will see your Kubernetes cluster as Active.

From here you can see a bunch of info about your new Kubernetes cluster. Also notice that you could even launch Kubectl right from hereand start running commands! Take some time to click around to see all the familiar stuff you are used to working with in Kubernetes. This is pretty cool and simplifies the management experience for Kubernetes. 

If you want to add more nodes or need the configuration code again just click the ellipsis button and edit.

In Edit Cluster you can change the cluster name, get and change settings and copy the code to add more VMs to the cluster.

That’s the end of this post. Thanks for reading. Check back for more Azure, Kubernetes, and Rancher blog posts.

Read More

Require Many Tags on Resource Groups via Azure Policy

Azure Policy can be used to enforce rules and effects on resources in your Azure subscriptions. It is a part of the Azure Governance and management toolbox native to Azure. I actually wrote a blog post all about Azure Policy here as a part of my native cloud management in azure blog series.

In this blog post I want to dig into Requiring Tags on Resource Groups via Azure Policy. There is a sample policy ARM Template to accomplish this here:

https://docs.microsoft.com/en-us/azure/governance/policy/samples/enforce-tag-on-resource-groups . What is not clear with the this policy template is how to add an enforce additional tags within the single ARM Template. This is important as you don’t want to have multiple templates to enforce multiple tags.

Well its actually pretty straight forward. You need to add the additional tags as Rules and Parameters. For example:

{
"not": {
"field": "[concat('tags[',parameters('Environment'), ']')]",
"exists": "true"
}
},

and

"Environment": {
"type": "String",
"metadata": {
"description": "Provides information on what the resource group is used for (useful for maintenance, policy enforcement, chargeback, etc.) Tag value: Dev, QA, Stage, Test, Prod. Example: Prod"
}
},

Be sure you add a parameter for every rule. Also in the example I gave I removed the “equals”: “[parameters(‘tagValue’)]” from the rules because I did not want to populate the tag value. I simply needed to require the tag and leave the value open for the person creating the resource to fill in. Here is the full example Policy ARM Template here:

{
  "mode": "all",
  "policyRule": {
    "if": {
      "allOf": [
        {
          "field": "type",
          "equals": "Microsoft.Resources/subscriptions/resourceGroups"
        },
        {
          "not": {
            "field": "[concat('tags[',parameters('BillTo'), ']')]",
            "exists": "true"
          }
        },
        {
            "not": {
              "field": "[concat('tags[',parameters('Functional Area'), ']')]",
              "exists": "true"
            }
          },
          {
            "not": {
              "field": "[concat('tags[',parameters('Environment'), ']')]",
              "exists": "true"
            }
          },
          {
            "not": {
              "field": "[concat('tags[',parameters('AppOwner'), ']')]",
              "exists": "true"
            }
          }
      ]
    },
    "then": {
      "effect": "deny"
    }
  },
  "parameters": {
    "BillTo": {
      "type": "String",
      "metadata": {
        "description": "Provides a charge code or cost center to attribute the bill for the resources too. Tag value: Cost Center. Example: team@domain.com"
      }
    },
    "Functional Area": {
        "type": "String",
        "metadata": {
          "description": "Provides information on department or team is responsible for administering/supporting the application. Tag value: Team name/email. Example: 1506548"
        }
      },
      "Environment": {
        "type": "String",
        "metadata": {
          "description": "Provides information on what the resource group is used for (useful for maintenance, policy enforcement, chargeback, etc.) Tag value: Dev, QA, Stage, Test, Prod. Example: Prod"
        }
      },
      "AppOwner": {
        "type": "String",
        "metadata": {
          "description": "The Business app owner to contact. Tag value: Business App owners’ email. Example: name@domain.com"
        }
      }
  }
}

After you create the Policy definition using the ARM template it will look like this:

When you assign the policy you will need to complete the parameters:

The next time someone deploys a resource group without the required tags in the subscription this policy is assigned to it will fail.

Thanks for reading!

Read More

Architect your Cloud with Azure Blueprints

Today as a part of the Azure Governance and management announcements at Microsoft Ignite 2018 Azure Blueprints Public Preview was announced. Azure Blueprints are a core part of the cloud governance and management story. They go hand and hand with Management Groups and will take the enterprise management story of Azure up a level. In this blog post I will take a deep dive into Azure Blueprints explaining what they are and give an example of how they can be used.

NOTE: This is a long blog post so I have also published this content as a whitepaper. The whitepaper PDF can be downloaded here.

BLUEPRINTS OVERVIEW

At a high-level Azure Blueprints help you meet organizational cloud standards, patterns, and requirements through governed subscriptions enabled and enforced by the grouping of artifacts like ARM templates, Azure policies, RBAC role assignments, and resource groups within a Blueprint.

Blueprints can be used to lay a cloud foundation, as cloud patterns, and group cloud governance frameworks. Blueprints are a one-click solution for deploying a cloud foundation, pattern, or governance framework to an Azure subscription. Think of an Azure Blueprint as re-usable design parameters for cloud that can be shared and used across an enterprise.

Azure architects typically map out and plan the many aspects of a cloud foundation for an organization such as access management, networking, storage, policy, security/compliance, naming conventions, tagging, monitoring, backup, locations, and more. Now Azure architects can step this designing a step further build these designs as Azure Blueprints and then apply them to subscriptions. The Blueprints give architects a way to orchestrate the deployment of grouped components to speed up the development and provisioning of new Azure environments ensuring they are meeting organizational compliance.

BLUEPRINTS ARE NOT AZURE POLICY

Azure policy is a service targeted to resource properties that exists or when being deployed with allow or explicit deny policies. It is used to ensure resources in an Azure subscription adhere to requirements and standards of an organization.

Azure policies can exist on their own or be a part of an Azure Blueprint. Blueprints do not replace Policy they are one of the Artifact types that make up a Blueprint.

THE MAKEUP OF A BLUEPRINT

Definition

A Blueprint consists of a Definition. The Definition is the design of what should be deployed it consists of the name of the Blueprint, the description and the Definition location. The Definition Location is the place in the Management Group hierarchy where this Blueprint Definition will be stored and determines the level assignment is allowed at. Currently you must have Contributor access to a Management Group to be able to save a Blueprint Definition to it. A Blueprint can be assigned at or below the Management Group it has set in its Definition Location. Here is a diagram to visualize Blueprint Assignment in relation to Management Group hierarchy:

Azure Blueprint Assignment

Artifacts

The Definition is where Blueprint Artifacts are added. As of right now the following is a list of the Artifact types:

  • Policy Assignments – Lets you add an Azure Policy. This can be a built-in or custom policy.
  • Role Assignments – Lets you add a user, app, or group and set the role. Only built-in roles are currently supported.
  • Azure Resource Manager templates – Lets you add an ARM Template. This does not let you import a parameters file. It does let you pre-set the parameters or set the parameters during assignment of the Blueprint.
  • Resource Groups – Lets you add a Resource Group to be created as a part of this Blueprint.

In my opinion the ARM Template artifact is the most impactful of the Blueprint artifact types because you can define such a variety of resources here. It opens the Blueprint to the power of ARM in general. Hopefully in the future we will see more scripting capability or the ability to load PowerShell scripts, runbooks, and or Functions.

There are two levels in the Artifacts. The first level is Subscription. The second level is Resource Group. Resource Group artifacts cannot be added to a Resource Group artifact. A Resource Group artifact can be created in a Subscription. An ARM Template artifact can only be created in a Resource Group artifact. A Policy Assignments or Role Assignments can be created at either the Subscription or Resource Group level.

Assignment

After a Blueprint has been built it needs to be applied. Applying a Blueprint is known as Blueprint assignment. The assignment is essentially the “what was deployed” for a Blueprint. This is how the artifacts are pushed out to Azure and used to track and audit deployments in Azure.

Sequencing

When the assignment of a Blueprint is processed the default order of resource creation is:

  • Role assignment artifacts at the Subscription level
  • Policy assignment artifacts at the Subscription level
  • Azure Resource Manager template artifacts at the Subscription level
  • Resource group artifacts and its child artifacts (role assignment, policy assignment, ARM Templates) at the Resource Group level

When a blueprint includes multiple Azure Resource Manager templates there may be a need to customize the sequencing order in which the Blueprint will deploy artifacts during assignment. You customize the artifact deployment sequence by deploying a Blueprint from an ARM Template declaring a dependency within it or declaring a dependency within an ARM Template artifact in the Blueprint. You declare a dependency using the dependsOn property in JSON. This essentially is a string array of artifact names.

Resource Locking

In cloud environments consistency is key. Naturally Azure Blueprints can also leverage resource locking in Azure. Blueprints have a Locking Mode. This Locking Mode can be applied to None or All Resources and is determined during the assignment of the Blueprint. The decision on cannot be changed later. If a locking state needs to be removed, then you must first remove the Blueprint assignment.

Some Blueprint artifacts create resources during assignment. These resources can have the following state:

  • Not Locked – Can be changed. Can be deleted.
  • Read Only – Can’t be changed and can’t be deleted.
  • Cannot Edit / Delete – Create, update, and delete within the resource group.

Artifacts that become Resource groups get the state of Cannot Edit / Delete automatically but you can create, update, and delete resources within them.

The high-level stages of an Azure Blueprint are Create it, assign it to a scope, and track it.

Anatomy of a Blueprint:

Azure Blueprint Anatomy

 

Blueprint does have a REST API. I am not covering the REST API in this blog post as I have not had the opportunity to spend much time working with it yet.

Now let’s look at building and assigning an Azure Blueprint.

BUILD A BLUEPRINT

Now I am going to give an example of building and using an Azure Blueprint in a cloud foundation mock scenario. In my mock scenario I have 3 Azure subscriptions. Each subscription should have a Core services Resource Group consisting of a core VNet with 3 subnets, an NSG for each subnet, and the web subnet should be ready for DMZ traffic. For the core VNet and any additional VNet added to the Core Services Resource Group I need network watcher deployed to it.

Each subscription also should have a core storage account and a blob storage that is ready for general storage needs. I want a tag applied to any Blueprint assignment labeling it with the assignment name, so it is easy to track. The last requirement I have is that I need the CloudOps team to automatically be owner of all core services resources. To accomplish all of this I created the following Blueprint:

Now let’s walk through the parts of creating and assigning the Blueprint. The first step is to create the Blueprint Definition.

In the basics step I give it a meaningful name and meaningful description. I set the Definition Location to the root of my Management groups. Doing this will allow me to assign this Blueprint to all 3 subscriptions in turn creating the core services RG in each subscription.

Next the Artifacts need to be added. Note that when adding an Artifact at the Subscription level you have these options as types:

The Resource Group Artifact type is only available at the subscription level and the ARM template Artifact type is only available at the Resource Group level. I added the Resource Group that the core networking and core storage will be deployed into.

Read More

Multiple Linux VM Deployment ARM Template

I looked for an existing ARM template that would create multiple Linux VM’s. I found only one that creates some in a scale set. The use case I was working with did not call for a scale set so I needed a different template.

I found a simple ARM template for creating multiple Windows VM’s on Azure here. It had exactly what I needed for my use case but did not cover Linux.

I modified the template and uploaded to Github in case this is helpful to anyone else. The repo has two templates. There is one for Ubuntu and one for SUSE. When you deploy the template it will need the following parameters:

The ARM template will create an availability set (AS) with N number of VM’s put in that AS, network interfaces, and public IP’s for each VM along with a VNet and Subnet as shown in the following screenshot:

Here is the link to download the ARM Template:

https://github.com/Buchatech/Multiple-Linux-VM-Deployment-ARM-Template

Read More