Website under construction

Figure 2-38:
Failover Cluster Manager console with two VMs running under Roles 
Figure 2-39:
Failover Cluster Manager console with three cluster shared volumes 
As you would expect, you can move all services and roles between all nodes in the cluster. It is a fully 
functional and updated cluster. This is a good time to take backups, especially of the cluster itself if 
you choose to restore prior to completing the upgrade process. 
At this point, you can begin the rolling operating system upgrades of the cluster. This is a test 
environment with few services running. In your environment, ensure that you have enough cluster 
resources to allow for one node at a time to be upgraded with the workloads supported by the 
remaining nodes active in the cluster. 
Begin by evicting one of the nodes from the cluster to start the rolling upgrade process. You can 
pause and evict the node from within the Failover Cluster Manager or, in Windows PowerShell, you 
can run the Suspend-ClusterNode cmdlet, followed by the Remove-ClusterNode cmdlet. You can 
choose any node in the cluster to begin the rolling upgrade. This example begins with ClusterHost01 
of the ContosoPVTCloud cluster. First, right-click the node and then, on the shortcut menu, select 
Pause to suspend it, as illustrated in Figure 2-40. 

---

41 
CHAPTER 2 | Software-defined datacenter 
Figure 2-40:
Pausing cluster node 
Next, right-click the node and select Evict to evict it, as depicted in Figure 2-41. 
Figure 2-41:
Evicting a cluster node 
Next, you can begin to install a new operating system installation on that node. 
Note You do not do an in-place operating system upgrade; you format and install a new operating 
system by using Windows Deployment Services, install from media, or other method of your choice 
to deploy the new operating system. 
After Windows Server 2016 has been installed on ClusterHost01, install the Hyper-V role, failover 
clustering, and, if necessary, Multipath I/O. Configure the network and storage as it was configured 
prior to the reinstallation. This is a good time to check for any updates available for this version of 
Windows Server 2016. Join ClusterHost01 to the ContosoCloud domain that contains the cluster being 
upgraded. 
Sign in to ClusterHost01 as a domain administrator of ContosoCloud domain or as another user with 
permissions on the current ContosoPVTCloud cluster. Start the Failover Cluster Manager and click Add 
Node, or run the Windows PowerShell Add-ClusterNode cmdlet. You must do this on the Windows 
Server 2016 server, not from a current cluster member running Windows Server 2012 R2. 
The Failover Cluster Manager console on ClusterHost01 shows that it has successfully joined the 
cluster, as shown in Figure 2-42. 

---

42 
CHAPTER 2 | Software-defined datacenter