When you upgrade the vSphere layer for the management components in the SDDC, you upgrade the management and compute Platform Services Controller and Management vCenter Server instances in Region A. You then repeat this operation in Region B.

Table 1. Upgrade Sequence fot the Management vCenter and Platform Services Controller Instances in the SDDC

Region

Upgrade Order

Role

IP Address

Fully Qualified Domain Name

Region A

1

Management Platform Services Controller that is configured in a highly available pair

172.16.11.61

sfo01m01psc01.sfo01.rainpole.local

2

Compute Platform Services Controller that is configured in a highly available pair

172.16.11.63

sfo01w01psc01.sfo01.rainpole.local

3

Management vCenter Server

172.16.11.62

sfo01m01vc01.sfo01.rainpole.local

Region B

4

Management Platform Services Controller that is configured in a highly available pair

172.17.11.61

lax01m01psc01.lax01.rainpole.local

5

Compute Platform Services Controller that is configured in a highly available pair

172.17.11.63

lax01w01psc01.lax01.rainpole.local

6

Management vCenter Server

172.17.11.62

lax01m01vc01.lax01.rainpole.local

Because the Platform Services Controller instances for the management and for the shared edge and compute clusters are configured in highly available pairs for each region, you follow a specific order when upgrading each component. You start the upgrade process by disabling traffic in the load balancer, upgrading the virtual appliance, then reenabling traffic in the load balancing for the Platform Services Controller instance for the management cluster. You then repeat this process for the Platform Services Controller instance for the shared edge and compute cluster.

Prerequisites

  • Download the vCenter Server Appliance Installer VMware-VCSA-all-6.7.x-build_number.iso file locally to a windows host that has access to the management VLAN network.

  • Review the release note of the vCenter Server version in this validated design.

  • Verify that all management ESXi hosts have lockdown mode disabled during the upgrade.

  • Verify that the management clusters do not have vSphere DRS set to Fully Automated.

  • Verify that any integration with the Management vCenter Server instances in the environment has been quiesced of all activities. Such activities include but are not limited to users performing active backups of components or provisioning of new virtual machines by using vRealize Automation. Without quiescing the environment, rollback operations can be disrupted by orphaned objects that might be generated after you have taken snapshots. You might also have to extend the time of the maintenance windows.

  • Verify that current backups of all Platform Services Controller and Management vCenter Server virtual appliances exist.