Create stretched cluster volumes and set up replication

Article
11/19/2024

Applies to: Azure Local, version 22H2

Important

Azure Stack HCI is now part of Azure Local. Product documentation renaming is in progress. However, older versions of Azure Stack HCI, for example 22H2 will continue to reference Azure Stack HCI and won't reflect the name change. Learn more.

This article describes how to create volumes and set up replication for stretched clusters in Azure Stack HCI using Windows Admin Center and PowerShell.

We'll create volumes on four servers in two sites, two servers per site as an example. Keep in mind however, that if you want to create three-way mirror volumes, you need at least six servers, three servers per site.

Stretched volumes and replication using Windows Admin Center

To create a volume and set up replication:

In Windows Admin Center, under Tools, select Volumes.
In the right pane, select the Inventory tab, then select Create.
In the Create volume panel, select Replicate volume between sites.
Select a replication direction between sites from the drop-down box.
Under Replication mode, select Asynchronous or Synchronous.
Enter a Source replication group name and a Destination replication group name.
Enter the desired size for the log volume.
Under Advanced, optionally do the following:
- Enter/change the Source replication group name.
- Enter/change the Destination replication group name.
- To use blocks already seeded on the target..., select that checkbox.
- To encrypt replication traffic, select that checkbox.
- To enable consistency groups, select that checkbox.
When finished, click Create.
In the right pane, verify that a data disk and a log disk are created in your primary (active) site, and that corresponding data and log replica disks are created in the secondary (passive) site. For bidirectional replication, you should see two sets of data and volume disks.
Under Tools, select Storage Replica.
In the right pane, under Partnerships, verify that the replication partnership has been successfully created.

Afterwards, you should verify successful data replication between sites before deploying VMs and other workloads. For more information, see the Verifying replication section in Validate the cluster.

Create stretched volumes using PowerShell

Volume creation is different for single-site standard clusters versus stretched (two-site) clusters. For both scenarios however, you use the New-Volume cmdlet to create a virtual disk, partition and format it, create a volume with matching name, and add it to cluster shared volumes (CSV).

Creating volumes and virtual disks for stretched clusters is a bit more involved than for single-site clusters. Stretched clusters require a minimum of four volumes - two data volumes and two log volumes, with a data/log volume pair residing in each site. Then you'll create a replication group for each site, and set up replication between them. We need to move resource groups around from server to server. The Move-ClusterGroup cmdlet is used to this.

First we move the Available Storage storage pool resource group to Server1 in Site1 using the Move-ClusterGroup cmdlet:
```
Move-ClusterGroup -Cluster ClusterS1 -Name 'Available Storage' -Node Server1
```

Next, create the first virtual disk (Disk1) for Server1 in Site1:

New-Volume -CimSession Server1 -FriendlyName Disk1 -FileSystem REFS -DriveLetter F -ResiliencySettingName Mirror -Size 10GB -StoragePoolFriendlyName "Storage Pool for Site 1"

Create a second virtual disk (Disk2) for Server1 in Site1:

New-Volume -CimSession Server1 -FriendlyName Disk2 -FileSystem REFS -DriveLetter G -ResiliencySettingName Mirror -Size 10GB -StoragePoolFriendlyName "Storage Pool for Site 1"

Now, take the Available Storage group offline:

Stop-ClusterGroup -Cluster ClusterS1 -Name 'Available Storage'

And move the Available Storage group to Server3 in Site2:

Move-ClusterGroup -Name 'Available Storage' -Node Server3

Create the first virtual disk (Disk3) on Server3 in Site2:

New-Volume -CimSession Server3 -FriendlyName Disk3 -FileSystem REFS -DriveLetter H -ResiliencySettingName Mirror -Size 10GB -StoragePoolFriendlyName "Storage Pool for Site 2"

And create a second virtual disk (Disk4) on Server3 in Site2:

New-Volume -CimSession Server3 -FriendlyName Disk4 -FileSystem REFS -DriveLetter I -ResiliencySettingName Mirror -Size 10GB -StoragePoolFriendlyName "Storage Pool for Site 2"

Now take the Available Storage group offline and then move it back to one of the servers in Site1:

Stop-ClusterGroup -Cluster ClusterS1 -Name 'Available Storage'

Move-ClusterGroup -Cluster ClusterS1 -Name 'Available Storage' -Node Server1

Using the Get-ClusterResource cmdlet, ensure that four virtual disk volumes were created, two in each storage pool:
```
Get-ClusterResource -Cluster ClusterS1
```

Now add Disk1 to Cluster Shared Volumes:

Add-ClusterSharedVolume -Name 'Cluster Virtual Disk (Disk1)'

You are done creating volumes, and ready to set up Storage Replica for replication.

Set up replication using PowerShell

When using PowerShell to set up Storage Replica for a stretched cluster, the disk that is used for the source data needs to be added as a Cluster Shared Volume (CSV). All other disks must remain as non-CSV drives in the Available Storage group. These disks are then added as Cluster Shared Volumes during the Storage Replica creation process.

In the previous step, the virtual disks were added using drive letters to make the identification of them easier. Storage Replica is a one-to-one replication, meaning a single disk can replicate to another single disk.

Step 1: Validate the topology for replication

Before starting, you should run the Test-SRTopology cmdlet for an extended period (like several hours). The Test-SRTopology cmdlet validates a potential replication partnership and validates the local host to the destination server or remotely between source and destination servers.

This cmdlet will verify that:

SMB can be accessed over the network, which means that TCP port 445 and port 5445 are open bi-directionally.
WS-MAN can be accessed over HTTP on the network, which means that TCP ports 5985 and 5986 are open.
An SR WMIv2 provider can be accessed and accepts requests.
Source and destination data volumes exist and are writable.
Source and destination log volumes exist with NTFS formatting or ReFS formatting and sufficient free space.
Storage is initialized in GPT format, not MBR, with matching sector sizes.
There's sufficient physical memory to run replication.

In addition, the Test-SRTopology cmdlet will also measure:

Round-trip latency of ICMP and report the average.
Performance counters for write Input/Output and report the average seen on that volume.
Estimated initial synchronization time.

Once Test-SRTopology completes, it will create an .html file (TestSrTopologyReport with date and time) in your Windows Temp folder. Any warning or failures should be reviewed as they could cause Storage Replica to not be properly created.

An example command that would run for 5 hours would be:

Test-SRTopology -SourceComputerName Server1 -SourceVolumeName W: -SourceLogVolumeName X: -DestinationComputerName Server3 -DestinationVolumeName Y: -DestinationLogVolumeName Z: -DurationInMinutes 300 -ResultPath c:\temp

Step 2: Create the replication partnership

Now that you completed the Test-SRTopology tests, you are ready to configure Storage Replica and create the replication partnership. In a nutshell, we'll configure Storage Replica by creating replication groups (RG) for each site and specifying the data volumes and log volumes for both the source server nodes in Site1 (Server1, Server2) and the destination (replicated) server nodes in Site2 (Server3, Server4).

Let's begin:

Add the Site1 data disk as a Cluster Shared Volume (CSV):

Add-ClusterSharedVolume -Name "Cluster Virtual Disk (Site1)"

The Available Storage group should be "owned" by the node it is currently sitting on. The group can be moved to Server1 using:
```
Move-ClusterGroup -Name "Available Storage" -Node Server1
```

To create the replication partnership, use the New-SRPartnership cmdlet. This cmdlet is also where you specify the source data volume and log volume names:

New-SRPartnership -SourceComputerName "Server1" -SourceRGName "Replication1" -SourceVolumeName "C:\ClusterStorage\Disk1\" -SourceLogVolumeName "G:" -DestinationComputerName "Server3" -DestinationRGName "Replication2" -DestinationVolumeName "H:" -DestinationLogVolumeName "I:"

The New-SRPartnership cmdlet creates a replication partnership between the two replication groups for the two sites. In this example Replication1 is the replication group for primary node Server1 in Site1, and Replication2 is the replication group for destination node Server3 in Site2.

Storage Replica will now set everything up. If there's any data to be replicated, it will do it here. Depending on the amount of data it needs to replicate, this may take a while. It's recommended to not move any groups around until this process completes.

Next steps

For related topics and other storage management tasks, see also:

Share via