In our lab environment with Big Data Cluster on VxRail HCI, we balanced services across the four nodes. The following figure provides a visual representation of how we distributed the Big Data Cluster services:

Figure 7.              Distribution of Big Data Cluster services

The pools used the nodes as follows:

• The SQL Server master pool used nodes 1, 2, and 4.
• The compute pool used nodes 1, 3, and 4.
• The data pool used nodes 1, 2, 3, and 4.
• The storage pool used nodes 2, 3, and 4.

The fourth node supported the external load balancer. The three SQL master VMs were used in an Always On availability group. SQL Server automatically builds out the Always On availability group when the HA option is selected during deployment. The availability group includes system databases and is integrated into the SQL Server engine so that any newly created databases are automatically added to the availabilty group. The Always On availability group provides the SQL master database architecture with several benefits:

• Protects the SQL master database by providing automatic and planned failover options
• Distributes read-intensive queries to one or more secondary replicas
• Enables backup operations to be offloaded to the secondary replicas
• Provides automatic page repair to protect against page corruption

An Always On cluster provides important features for production implementations of Big Data Cluster, such as high availability, offload support, and self-healing capabilities.

The Big Data Cluster configuration on the VxRail system performed as expected with no challenges or obstacles. Offering easy deployment and integration with Kubernetes and VMware, VxRail HCI is an ideal platform for a Big Data Cluster environment.