Deploy an Azure Stack HCI, version 23H2 system using the Azure portal
Applies to: Azure Stack HCI, version 23H2
This article helps you deploy an Azure Stack HCI, version 23H2 system using the Azure portal.
Prerequisites
- Completion of Register your servers with Azure Arc and assign deployment permissions.
- For three-node clusters, the network adapters that carry the in-cluster storage traffic must be connected to a network switch. Deploying three-node clusters with storage network adapters that are directly connected to each server without a switch isn't supported in this preview.
Start the wizard and fill out the basics
Open a web browser, navigate to Azure portal. Search for Azure Arc. Select Azure Arc and then go to Infrastructure | Azure Stack HCI. On the Get started tab, select Deploy cluster.
Select the Subscription and Resource group in which to store this system's resources.
All resources in the Azure subscription are billed together.
Enter the Cluster name used for this Azure Stack HCI system when Active Directory Domain Services (AD DS) was prepared for this deployment.
Select the Region to store this system's Azure resources. For a list of supported Azure regions, Azure requirements.
We don't transfer a lot of data so it's OK if the region isn't close.
Create an empty Key vault to securely store secrets for this system, such as cryptographic keys, local admin credentials, and BitLocker recovery keys.
Key Vault adds cost in addition to the Azure Stack HCI subscription. For details, see Key Vault Pricing.
Select the server or servers that make up this Azure Stack HCI system.
Important
Servers must not be joined to Active Directory before deployment.
Select Validate, wait for the green validation checkbox to appear, and then select Next: Configuration.
The validation process checks that each server is running the same exact version of the OS, has the correct Azure extensions, and has matching (symmetrical) network adapters.
Specify the deployment settings
Choose whether to create a new configuration for this system or to load deployment settings from a template–either way you'll be able to review the settings before you deploy:
Choose the source of the deployment settings:
- New configuration - Specify all of the settings to deploy this system.
- Template spec - Load the settings to deploy this system from a template spec stored in your Azure subscription.
- Quickstart template - This setting isn't available in this release.
Select Next: Networking.
Specify network settings
For multi-node clusters, select whether the cluster is cabled to use a network switch for the storage network traffic:
- No switch for storage - For two-node clusters with storage network adapters that connect the two servers directly without going through a switch.
- Network switch for storage traffic - For clusters with storage network adapters connected to a network switch. This also applies to clusters that use converged network adapters that carry all traffic types including storage.
Choose traffic types to group together on a set of network adapters–and which types to keep physically isolated on their own adapters.
There are three types of traffic we're configuring:
- Management traffic between this system, your management PC, and Azure; also Storage Replica traffic
- Compute traffic to or from VMs and containers on this system
- Storage (SMB) traffic between servers in a multi-node cluster
Select how you intend to group the traffic:
Group all traffic - If you're using network switches for storage traffic you can group all traffic types together on a set of network adapters.
Group management and compute traffic - This groups management and compute traffic together on one set of adapters while keeping storage traffic isolated on dedicated high-speed adapters.
Group compute and storage traffic - If you're using network switches for storage traffic, you can group compute and storage traffic together on your high-speed adapters while keeping management traffic isolated on another set of adapters.
This is commonly used for private multi-access edge compute (MEC) systems.
Custom configuration - This lets you group traffic differently, such as carrying each traffic type on its own set of adapters.
Tip
If you're deploying a single server that you plan to add servers to later, select the network traffic groupings you want for the eventual cluster. Then when you add servers they automatically get the appropriate settings.
For each group of traffic types (known as an intent), select at least one unused network adapter (but probably at least two matching adapters for redundancy).
Make sure to use high-speed adapters for the intent that includes storage traffic.
For the storage intent, enter the VLAN ID set on the network switches used for each storage network.
Important
Portal deployment does not allow to specify your own IPs for the storage intent. However, you can use ARM template deployment if you require to specify the IPs for storage and you cannot use the default values from Network ATC. For more information check this page: Custom IPs for storage intent
To customize network settings for an intent, select Customize network settings and provide the following information:
- Storage traffic priority. This specifies the Priority Flow Control where Data Center Bridging (DCB) is used.
- Cluster traffic priority.
- Storage traffic bandwidth reservation. This parameter defines the bandwidth allocation in % for the storage traffic.
- Adpater properties such as Jumbo frame size (in bytes) and RDMA protocol (which can now be disabled).
Using the Starting IP and Ending IP (and related) fields, allocate a contiguous block of at least six static IP addresses on your management network's subnet, omitting addresses already used by the servers.
These IPs are used by Azure Stack HCI and internal infrastructure (Arc Resource Bridge) that's required for Arc VM management and AKS Hybrid.
Select Next: Management.
Specify management settings
Optionally edit the suggested Custom location name that helps users identify this system when creating resources such as VMs on it.
Select an existing Storage account or create a new Storage account to store the cluster witness file.
When selecting an existing account, the dropdown list filters to display only the storage accounts contained in the specified resource group for deployment. You can use the same storage account with multiple clusters; each witness uses less than a kilobyte of storage.
Enter the Active Directory Domain you're deploying this system into.
This must be the same fully qualified domain name (FQDN) used when the Active Directory Domain Services (AD DS) domain was prepared for deployment.
Enter the OU created for this deployment. For example:
OU=HCI01,DC=contoso,DC=com
Enter the Deployment account credentials.
This domain user account was created when the domain was prepared for deployment.
Enter the Local administrator credentials for the servers.
The credentials must be identical on all servers in the system. If the current password doesn't meet the complexity requirements (12+ characters long, a lowercase and uppercase character, a numeral, and a special character), you must change it on all servers before proceeding.
Select Next: Security.
Set the security level
Select the security level for your system's infrastructure:
- Recommended security settings - Sets the highest security settings.
- Customized security settings - Lets you turn off security settings.
Select Next: Advanced.
Optionally change advanced settings and apply tags
Choose whether to create volumes for workloads now, saving time creating volumes and storage paths for VM images. You can create more volumes later.
Create workload volumes and required infrastructure volumes (Recommended) - Creates one thinly provisioned volume and storage path per server for workloads to use. This is in addition to the required one infrastructure volume per cluster.
Create required infrastructure volumes only - Creates only the required one infrastructure volume per cluster. You'll need to later create workload volumes and storage paths.
Use existing data drives (single servers only) - Preserves existing data drives that contain a Storage Spaces pool and volumes.
To use this option you must be using a single server and have already created a Storage Spaces pool on the data drives. You also might need to later create an infrastructure volume and a workload volume and storage path if you don't already have them.
Important
Don't delete the infrastructure volumes created during deployment.
Here's a summary of the volumes that are created based on the number of servers in your system. To change the resiliency setting of the workload volumes, delete them and recreate them, being careful not to delete the infrastructure volumes.
# Servers Volume resiliency # Infrastructure volumes # Workload volumes Single server Two-way mirror 1 1 Two servers Two-way mirror 1 2 Three servers + Three-way mirror 1 1 per server Select Next: Tags.
Optionally add a tag to the Azure Stack HCI resource in Azure.
Tags are name/value pairs you can use to categorize resources. You can then view consolidated billing for all resources with a given tag.
Select Next: Validation. Select Start validation.
The validation will take about 15 minutes for one to two server deployment and more for bigger deployments. Monitor the validation progress.
Validate and deploy the system
After the validation is complete, review the validation results.
If the validation has erorrs, resolve any actionable issues, and then select Next: Review + create.
Don't select Try again while validation tasks are running as doing so can provide inaccurate results in this release.
Review the settings that will be used for deployment and then select Review + create to deploy the system.
The Deployments page then appears, which you can use to monitor the deployment progress.
If the progress doesn't appear, wait for a few minutes and then select Refresh. This page may show up as blank for an extended period of time owing to an issue in this release, but the deployment is still running if no errors show up.
Once the deployment starts, the first step in the deployment: Begin cloud deployment can take 45-60 minutes to complete. The total deployment time for a single server is around 1.5-2 hours while a two-node cluster takes about 2.5 hours to deploy.
Verify a successful deployment
To confirm that the system and all of its Azure resources were successfully deployed
In the Azure portal, navigate to the resource group into which you deployed the system.
On the Overview > Resources, you should see the following:
Number of resources Resource type 1 per server Machine - Azure Arc 1 Azure Stack HCI 1 Arc Resource Bridge 1 Key vault 1 Custom location 2* Storage account 1 per workload volume Azure Stack HCI storage path - Azure Arc * One storage account is created for the cloud witness and one for key vault audit logs. These accounts are locally redundant storage (LRS) account with a lock placed on them.
Rerun deployment
If your deployment fails, you can rerun the deployment. In your cluster, go to Deployments and in the right-pane, select Rerun deployment.
Post deployment tasks
After the deployment is complete, you may need to perform some additional tasks to secure your system and ensure it's ready for workloads.
Enable RDP
For security reasons, Remote Desktop Protocol (RDP) is disabled and the local administrator renamed after the deployment completes on Azure Stack HCI systems. For more information on the renamed administrator, go to Local builtin user accounts.
You may need to connect to the system via RDP to deploy workloads. Follow these steps to connect to your cluster via the Remote PowerShell and then enable RDP:
Run PowerShell as administrator on your management PC.
Connect to your Azure Stack HCI system via a remote PowerShell session.
$ip="<IP address of the Azure Stack HCI server>" Enter-PSSession -ComputerName $ip -Credential get-Credential
Enable RDP.
Enable-ASRemoteDesktop
Note
As per the security best practices, keep the RDP access disabled when not needed.
Disable RDP.
Disable-ASRemoteDesktop
Lock Arc Resource bridge
The Arc Resource Bridge enables the Azure Arc services to manage your Azure Stack HCI system. To prevent the accidental deletion, we recommend that you lock the Arc Resource Bridge resource. Follow these steps to configure the resource locks:
- In the Azure portal, navigate to the resource group into which you deployed your Azure Stack HCI system.
- On the Overview > Resources tab, you should see an Arc Resource Bridge resource.
- Select and go to the resource. In the left pane, select Locks. To lock the Arc Resource Bridge, you must have the Azure Stack HCI Administrator role for the resource group.
- In the right pane, select Add.
- Specify a Lock name. Make sure to choose Delete as Lock type for the Arc Resource Bridge, and then select OK.
Note
Remove the lock on the Arc Resource Bridge before you apply any solution updates.
For more information, see Configure locks to prevent accidental deletion.
Next steps
- If you didn't create workload volumes during deployment, create workload volumes and storage paths for each volume. For details, see Create volumes on Azure Stack HCI and Windows Server clusters and Create storage path for Azure Stack HCI.
- Get support for Azure Stack HCI deployment issues.