Three or more LARGE platform bricks can be connected together to form a platform cluster.

Note:

Ensure that you take a backup of the Platform1 node before you create clusters. Refer to VMware best practices to take the backup of virtual machines (like VMware VDP using VADP). Restore the Platform1 node from backup if there is an unrecoverable error while creating the cluster. It is recommended that you use cleanly deployed platform nodes while creating clusters. Redeploy the new platform nodes (p2-pn) before restarting cluster creation process if there is an unrecoverable error.

To decide the required number of platform bricks:

Number of bricks needed = Round off to next Integer ((Total number of managed VMs) / (Capacity of LARGE Platform brick in table above))

Scaling up Scenarios for the Platform Cluster

  • Scenario 1

    1. Assume that on January 1st (today), the datacenter has 2000 VMs (with flows) across many vCenters.

    2. Assume that in March, the number of VMs grows to 3100.

    3. Assume that in June, the number of VMs grows to 6100 which could be because of the additions of few more vCenters or the expansion of the existing vCenters.

    4. Assume that in December, the number of VMs grows to 18100 (with flows).

    The deployment model for this scenario is as follows:

    1. On January 1, deploy a single platform node with the MEDIUM brick size.

    2. In March, scale up the platform node to the LARGE brick size.

    3. In June, scale out the platform, convert to a three node platform cluster by adding new Platform nodes to the existing Platform.

    4. In December, the user needs a four node platform cluster.

  • Scenario 2

    1. Assume that on January 1st (today), the datacenter has 7000 VMs (with flows) across many vCenters.

    2. Assume that in June, the number of VMs grows to 15000 (with flows).

    3. Assume that in December, the number of VMs grows to 24000 (with flows).

    The deployment model for this scenario is as follows:

    1. On January 1, deploy a three node platform cluster.

    2. In June or later, as the environment size gets closer to exceeding 18000, the user needs a four node platform cluster.

    3. In December, as the environment size gets closer to exceeding 24000, the user needs a five node platform cluster.