Next, we created four VMs—one on each of the PowerFlex R840 HCI nodes—for the Hadoop cluster consisting of three virtual machines and the Oracle NoSQL application. We then created two VMs—one on each of the compute nodes—for the Oracle and Microsoft SQL Server database applications.

We used the following design principles and best practices to create the VMs:

• Hard disk drives—We used the following disk types for all database-related virtual disks:

• Raw Device Mapping (RDM)—Disks used in Oracle deployment are presented as raw devices to the database VM.
• Using Virtual Machine File System (VMFS) datastores—We created hard disks using VMFS datastores for all VMs except Oracle. This includes VMs for Hadoop name node, Hadoop data node 1, Hadoop data node 2, Oracle NoSQL, and Microsoft SQL Server.

• VM vCPU and vMem—The following table shows the amount of virtual CPU (vCPU) and virtual memory (vMem) that we assigned to the application VMs:

Table 5.            VM configuration: vCPU and vMem details

• Enable disk UUID—In each of the VM options, we added the disk.enableUUID parameter and set it to TRUE. This setting ensures that the Virtual Machine Disk, (VMDK) always presents a consistent disk UUID to the VM.
• We used different SCSI controllers for operating system disks and other (database application) disks. We assigned SCSI controller 0 to the operating system disks, and SCSI controller 1 to the other disks.