Step by step Hyper-V failover cluster migration to Windows Server 2012 R2
This guide is a step by step process to migrate a Hyper-V failover cluster to Windows Server 2012 R2. The guide demonstrates migrating from Windows Server 2008 R2, but the same guide can be applied when migrating from Windows Server 2012.
The scope of this migration covers the same basic setup that has been outlined in my earlier Hyper-V cluster setup guide starting here: http://bit.ly/1aRxST2 Our hardware consists of two DL360p servers, connected to an HP StorageWorks P2000 by means of iSCSI. Our migration will follow the path of evicting a node from the existing cluster, formatting it, and bringing up a new cluster on Windows Server 2012 R2 to migrate to.
Step 1: Verification
Before beginning anything else, you want to ensure that your current infrastructure is running healthy. At this step you want to make sure that there are no underlying hardware and/or software issues that are going to cause you any problems during the migration. We will be dismantling our redundancy during the migration, albeit temporarily, but you want to be sure you can run all workloads on the single host without issue. Also make sure that your backup solutions are in place and working properly.
Step 2: Evict one node
The goal of this step is to move everything off one of the nodes so we can be in the position to take the node out of the cluster and prepare it for Windows Server 2012 R2. So, on your source cluster you want to move all your cluster services onto a single node. In this example, I will utilize live migration to move the virtual machines:
After you complete the migration of everything over to a single node, we can proceed with evicting the other node from the cluster. I like to shut off the server that will be evicted at this point, before evicting. This allows me to verify that everything is running ok on the single node before committing any changes. To evict the node, on the source cluster within failover cluster expand nodes and right click the node to evict (note the down arrow corresponding to the node that us currently down). Under more actions, choose the evict option:
Step 3: Prepare destination cluster
Now that we have freed up one server, we are going to use this server to prepare the destination cluster on Windows Server 2012 R2. The setup for this follows my previous blog entry for configuring a Hyper-V cluster, I am not going to go into too much detail in the actual cluster setup here. Setup the newly formatted server just as you were going to configure a Windows Server 2012 R2 failover cluster from scratch. If you need guidance with this step, check out my blog entries on configuring this here: http://bit.ly/1aRxST2
When you get your new Windows Server 2012 R2 failover cluster up and running there will be no storage or roles. This is because these are still dedicated to the old cluster. We are going to use the Copy Cluster Roles feature to migrate these to the new cluster. To do this, fire up failover cluster manager on your new Windows Server 2012 R2 node, right click the cluster name, and choose More Actions - Copy Cluster roles:
This will launch the copy cluster roles wizard:
Specify the name of the old cluster in the first step of the wizard
When you hit next the wizard will examine all the roles that can be copied over. For a more detailed report, click View Report. The report shown will show more details about the roles that can be copied over, including any that are not eligible to be copied:
In the final step of this wizard you are asked for confirmation. Once you hit next, the wizard goes through the process of copying the roles over to the new cluster and presents you with a report of the results:
This process has not actually migrated any of the storage over to the new cluster yet. If you look at the roles, and storage it will actually show as being offline on the new cluster.
Step 4: Migration
This is the first part where you will require a maintenance window to complete this task. During this step, we will bring down all resources on the old cluster, and bring them back online on the new cluster. The length of time you will need to complete this step varies based on how many VMs are running within your cluster.
So for the first part, what we want to do is shut down each and every role that is located on the old cluster. You can do this by logging into the console of each VM and shutting it down, or right clicking each VM and choosing the shut down option within failover cluster manager. You don’t want to risk any data corruption when you take down the cluster shared volume:
Next, take down the cluster shared volume. This is done within failover cluster manager, under cluster shared volumes. Right click the disk and choose the option to take offline:
Once the disk and all VM’s are offline, we can jump over to the new cluster and bring them back online. Start with the disk first: Jump into failover cluster manager on the new server, under storage and disks, and right click your cluster shared volume, choosing the option bring online:
Once the disk is online, you can bring your VM’s up in failover cluster manager under roles, right clicking each and choosing the option start:
You will probably want to upgrade the Hyper-V integration services on each VM at this point as well.
Step 5: Retire old cluster
When you have verified everything is up and running on the new cluster, we can proceed with decommissioning the old cluster to free up the server and get it added to the new cluster. What we want to do now is remove all services (VMs) from the old cluster. On the old server, fire up failover cluster manager and right click each VM choosing the option delete for each one:
And finally, once all services have been removed, right click the cluster name and choose the option More Actions - “Destroy cluster”:
Step 6: Finish up
Finally, we are going to take our old cluster node and bring it back into the new cluster. Again, follow my steps in the Hyper-V failover cluster setup to get your server prepared to the point at which you would create the cluster. Instead of creating a cluster however, we are going to add the node to the existing cluster. We do this in failover cluster manager by right clicking our cluster name and choosing the option “Add Node…”
This will launch another wizard that will walk you through adding the node to the cluster. During the wizard, specify the name of the additional node to be added
The next step will run a validation on your cluster. The default option will run all tests, including storage tests. When the storage tests are ran, the storage is temporarily taken offline to complete the test. Therefore, before running the test, ensure you have a maintenance window to take your workloads down and make sure you offline your VMs. You can choose to skip the tests if you want, but its recommended you don’t.
Hopefully at this point your validation goes well. If not, review the report and make adjustments where necessary. If everything goes well, your node will be added to your cluster:
The final step that I do is configure the cluster quorum settings. Since we are dealing with a 2 node cluster, we will require a disk witness. To configure your quorum settings, right click your cluster name and choose More Actions - Configure Cluster Quorum Settings:
And finally choose the option to use the default quorum configuration:
That should do it! If you made it this far, congrats!