Step-by-Step Configuring Amazon EC2 Business-Critical Applications Part 2

Part 1 of this series covered the steps needed to prepare Amazon Elastic Compute Cloud’s (EC2) infrastructure for support of a Windows Server Failover Cluster, (WSFC), that spans availability zones. The steps involved in preparing the virtual private cloud (VPC), storage, and enabling core WSFC components were all covered.
This article will continue where we left off, and we’ll continue with the installation of and configuration for a SQL Server 2019 failover instance (FCI).
Cluster Storage
A shared storage device is required for a SQL Server FCI. A SQL Server FCI that is built on-premises will usually store its data on a SAN. All cluster nodes have access to the same LUN on SAN. WSFC controls access through SCSI 3 reservations to ensure that only the active cluster node can access the storage. WSFC ensures that a failover occurs and that the new node comes online has access to the storage. All other nodes are locked out.
A new option for SQL Server FCI storage has been added to Windows Server 2012 and SQL Server 2012. Windows Server 2012 introduced SMB 3. It included SMB Transparent Failover and significant performance enhancements. These new features enabled SMB file servers to be used in place of SQL Server FCI storage.
Third-party software solutions, such as SIOS DataKeeper Cluster Edition, allow users to create SANless clusters. Instead of using block-level replication, they can use block-level replication to keep cluster data in sync among the nodes. Cluster configurations can be created that use locally attached storage by eliminating the need for a SAN. This allows you to create clusters that span geographical locations. Instead of controlling SCSI reservations the SANless cluster solution controls replication direction. This ensures that the active cluster node is always at the source of replication and the remaining nodes are at the target.
Part 1 explains that you must create a cluster that spans multiple AZs in order to be eligible for Amazon Web Services’ 99.99% availability SLA for your compute resource resources. To meet the 99.99 percent application availability SLA, in this example, we will use Amazon FSx, which is a hosted SMB 3 service you can use to store cluster data, and SIOS DataKeeper Cluster edition. Amazon FSx offers a 99.9 per cent availability guarantee. This becomes the SLA for your entire cluster.
Configuring the Cluster Storage
Start the DataKeeperUI on one of the cluster nodes, and create your DataKeeper volume resource as shown below.
Figure 1 Connect to both servers — SQL1 first, then SQL2. The Server Overview Report should look like this if you have connected to both servers.
[Click on the image to see a larger version.] Figure 2 Click on Create Job to launch the Job Creation Wizard.
[Click on the image to see a larger view.] Figure 3 DataKeeper supports bothsynchronous and asynchronous replication. Choose synchronous to replicate between AZs within the same region. Asynchronous is best if you need to replicate across regions and cloud providers. To register the DataKeeper Volume resource in the cluster’s Available Storage, click “Yes” at the prompt.
Figure 4 DataKeeper Volume D is now visible in Failover Cluster manager under Available Storage.
[Click on the image to see a larger version.] Figure 5 Install the First Node of SQL Server FCI from SQL1
Now that the core cluster is created and the DataKeeper volume resource has been allocated, it’s time to install SQL Server at the first cluster node. The following example shows a cluster configuration using SQL Server 2019, Win, as mentioned earlier.