Plan an Azure Arc-enabled data services deployment

This article describes how to plan to deploy Azure Arc-enabled data services.

Deployment steps

In order to experience Azure Arc-enabled data services, you'll need to complete the following tasks.

1. Plan your deployment

   The details in this article will guide your plan.

2. Install client tools.

3. Register the Microsoft.AzureArcData provider for the subscription where the Azure Arc-enabled data services will be deployed, as follows:

   az provider register --namespace Microsoft.AzureArcData
   

4. Access a Kubernetes cluster.

   For demonstration, testing, and validation purposes, you can use an Azure Kubernetes Service cluster. To create a cluster, follow the instructions at Quickstart: Deploy Azure Arc-enabled data services - directly connected mode - Azure portal to walk through the entire process.

5. Create Azure Arc data controller in direct connectivity mode (prerequisites).

   For other ways to create a data controller see the links under Related content.

6. Create data services.

   For example, Create a SQL Managed Instance enabled by Azure Arc.

7. Connect with Azure Data Studio.

As you begin planning to deploy Azure Arc-enabled data services, it's important to properly understand your database workloads and your business requirements for those workloads. For example, you need to consider availability, business continuity, and capacity requirements for memory, CPU, and storage for the workloads. And you need to carefully prepare the infrastructure to support the database workloads, based on your business requirements.

Prerequisites

Before you start, be sure that you've met certain prerequisites and have the necessary background and information ready. To ensure a successful deployment, your infrastructure environment must be properly configured with the right level of access and the appropriate capacity for storage, CPU, and memory.

Review the following articles:

• Sizing guidance
• Storage configuration
• Connectivity modes and their requirements

Verify that:

• The arcdata CLI extension is installed.

• The other client tools are installed.

• You have access to the Kubernetes cluster.

• Your kubeconfig file is configured. It should point to the Kubernetes cluster that you want to deploy to. To verify the current context of the cluster, run the following command:

  kubectl cluster-info
  

• You have an Azure subscription that resources such as an Azure Arc data controller, SQL Managed Instance enabled by Azure Arc, or Azure Arc-enabled PostgreSQL server will be projected and billed to.

• The Microsoft.AzureArcData provider is registered for the subscription where the Azure Arc-enabled data services will be deployed.

After you're prepared the infrastructure, deploy Azure Arc-enabled data services in the following way:

1. Create an Azure Arc-enabled data controller on one of the validated distributions of a Kubernetes cluster.
2. Create a SQL Managed Instance enabled by Azure Arc and/or an Azure Arc-enabled PostgreSQL server.

Deployment requirements

You can deploy Azure Arc-enabled data services on various types of Kubernetes clusters. Currently, the validated list of Kubernetes services and distributions includes:

• Amazon Elastic Kubernetes Service (Amazon EKS)
• Azure Kubernetes Service (AKS)
• Azure Kubernetes Service on Azure Stack HCI
• Azure Red Hat OpenShift
• Google Kubernetes Engine (GKE)
• Open source, upstream Kubernetes (typically deployed by using kubeadm)
• OpenShift Container Platform (OCP)
• K3s
• Additional partner-validated Kubernetes distributions

Deployment information

When you're creating Azure Arc-enabled data services, regardless of the service or distribution option you choose, you'll need to provide the following information:

• Data controller name: A descriptive name for your data controller (for example, production-dc or seattle-dc). The name must meet Kubernetes naming standards.

• Username: The username for the Kibana/Grafana administrator user.

• Password: The password for the Kibana/Grafana administrator user.

• Name of your Kubernetes namespace: The name of the Kubernetes namespace where you want to create the data controller.

• Connectivity mode: Determines the degree of connectivity from your Azure Arc-enabled data services environment to Azure. Your choice of connectivity mode determines the options for deployment methods. For more information, see Connectivity modes and requirements.

• Azure subscription ID: The Azure subscription GUID for where you want to create the data controller resource in Azure. All deployments of SQL Managed Instance enabled by Azure Arc and Azure Arc-enabled PostgreSQL are also created in and billed to this subscription.

• Azure resource group name: The name of the resource group where you want to create the data controller resource in Azure. All deployments of SQL Managed Instance enabled by Azure Arc and Azure Arc-enabled PostgreSQL are also created in this resource group.

• Azure location: The Azure location where the data controller resource metadata will be stored in Azure. For a list of available regions, see the Products available by region page for Azure global infrastructure. The metadata and billing information about the Azure resources that are managed by your deployed data controller is stored only in the location in Azure that you specify as the location parameter. If you're deploying in direct connectivity mode, the location parameter for the data controller is the same as the location of your targeted custom location resource.

• Service principal information:

  • If you're deploying in indirect connectivity mode, you'll need service principal information to upload usage and metrics data. For more information, see the "Assign roles to the service principal" section of Upload usage data, metrics, and logs to Azure.

• Infrastructure: For billing purposes, you must indicate the infrastructure on which you're running Azure Arc-enabled data services. The options are:

• alibaba

• aws

• azure

• gcp

• onpremises

• other

• Container runtime: Use containerd runtime for the container runtime. Azure Arc-enabled data services don't support Docker runtime.

Additional concepts for direct connectivity mode

As outlined in Connectivity modes and requirements, you can deploy the Azure Arc data controller either in direct or indirect connectivity mode. Deploying Azure Arc data services in direct connectivity mode requires additional concepts and considerations:

• First, the Kubernetes cluster where the Azure Arc-enabled data services will be deployed needs to be an Azure Arc-enabled Kubernetes cluster. By connecting your Kubernetes cluster to Azure, you can deploy and manage Azure Arc data services to your cluster directly from the Azure portal, upload your usage, logs and metrics to Azure automatically and get several other Azure benefits. To learn how, see Connect your cluster to Azure.

• After the Kubernetes cluster is Azure Arc-enabled, deploy Azure Arc-enabled data services by doing the following:

  1. Create the Azure Arc data services extension. To learn how, see Cluster extensions on Azure Arc-enabled Kubernetes.
  2. Create a custom location. To learn how, see Custom locations on top of Azure Arc-enabled Kubernetes.
  3. Create the Azure Arc data controller.

  You can perform all three of these steps in a single step by using the Azure Arc data controller creation wizard in the Azure portal.

After you've installed the Azure Arc data controller, you can create and access data services such as SQL Managed Instance enabled by Azure Arc or Azure Arc-enabled PostgreSQL server.

Known limitations

Currently, only one Azure Arc data controller per Kubernetes cluster is supported. However, you can create multiple Arc data services, such as Arc-enabled SQL managed instances and Arc-enabled PostgreSQL servers, that are managed by the same Azure Arc data controller.

Related content

You have several additional options for creating the Azure Arc data controller:

Just want to try things out? Get started quickly with Azure Arc Jumpstart on AKS, Amazon EKS, or GKE, or in an Azure VM.

• Create a data controller in direct connectivity mode with the Azure portal
• Create a data controller in indirect connectivity mode with CLI
• Create a data controller in indirect connectivity mode with Azure Data Studio
• Create a data controller in indirect connectivity mode from the Azure portal via a Jupyter notebook in Azure Data Studio
• Create a data controller in indirect connectivity mode with Kubernetes tools