Upgrading a highly available VMM 2012 R2 to VMM 2016

There are some differences between upgrading standalone VMM and highly available. Since VMM 2016, you no longer need to remove a cluster and create a new one. Using a Cluster OS Rolling Upgrade and its mixed mode, you can upgrade VMM nodes one by one from Windows Server 2012 R2 to Windows Server 2016, and then upgrade VMM 2012 R2 to 2016. See the Cluster OS Rolling Upgrade recipe in Chapter 9, Managing Clouds, Fabric Up-dates, Resources, Clusters and new Features of 2016 for details.

1. Perform a backup of the current VMM database. You can use the steps from the Upgrading VMM 2012 R2 to VMM 2016 section.

1. If your VMM is configured with AlwaysOn AGs, you need to remove the VMM database from AG before upgrading the VMM. For doing this, open SQL Server Management Studio, connect to the server and expand AlwaysOn High Availability, select VirtualManagerDB under Availability Databases, right-click, and then choose Remove Database from Availability Group...:

1. Initiate manual failover to the SQL Server on which the VMM database is installed:

1. Identify the active and passive VMM nodes. For doing this, open Failover Cluster Manager (cluadmin.msc), click on Roles and note Owner Node for VMM HA role, as shown in the following screenshot. VMM02, in this example, is the active node:

1. Uninstall VMM 2012 R2 from the passive node, retaining the database.
2. Place the passive node into maintenance mode. For doing this, open Failover Cluster Manager (cluadmin.msc), click on Nodes and right-click on the passive node, then choose Pause and Drain Roles, then choose More Actions - Evict to remove the passive node from the cluster.

1. Upgrade OS to Windows Server 2016 on the passive node, and then add the node back to the cluster (Add-ClusterNode).
2. Install VMM 2016, confirm high available configuration and upgrade the existing database when prompted.
3. Do failover to the passive node using Failover Cluster Manager.
4. Repeat steps 5-8 for the other nodes. Once you've upgraded all the nodes, update the cluster functional level by running the Update-ClusterFunctionalLevel command.
5. After upgrading, you will need to configure AlwaysOn AG (if any) for the VMM database. Refer to the Configure the VMM database with AlwaysOn AGs recipe in Chapter 4, Installing a Highly Available VMM Server.