Windows clusters have always presented a special problem for those tasked with patch management. On one hand, clustering is very necessary. Clustering provides fault tolerance and is often the only way to protect a mission critical server or service. On the other hand, failover clustering compounds the difficulty of patch management.
Cluster Patching – The Old Way
Prior to the release of Windows Server 2012, if you wanted to patch a cluster (using only native tools) you had to move the clustered resources off of a cluster node, patch and reboot the cluster node, and then repeat the process for the other nodes in the cluster. The technique worked, but it required manual intervention and was tedious and time consuming. The manual nature of the process could be forgiven if patching was a one time process, but as any good admin knows, patching is an ongoing process.
Cluster Patching – The New Way
Those cluster administrators who dread Patch Tuesday will be happy to know that you don’t have to use a manual process to patch a Windows Server 2012 cluster. In fact, Microsoft has introduced a new feature called Cluster Aware Updating.
The tricky part of patching a Windows Server 2012 cluster is that cluster aware updating is not used by default. You have to enable cluster aware updating. Otherwise, the cluster will have to be patched using the same manual technique required by clusters running earlier versions of Windows Server.
How Does Cluster Aware Patching Work?
Before I explain how to enable cluster aware updating, you might be curious as to how the patching process works in a Windows Server 2012 cluster. Cluster aware updating works similarly to the method used to manually patch a failover cluster, except that the process is automated.
Windows Server 2012 uses a round robin approach to updating cluster nodes. The process begins by identifying the cluster node that has the most free memory available. The clustered resources are moved from a random cluster node to the node that has the most free memory. After the clustered resources have been moved off of the cluster node, the node is placed into maintenance mode. At this point, the cluster node is patched, rebooted, and then taken out of maintenance mode. The process is then repeated for every remaining node in the cluster.
How Is Cluster Aware Patching Implemented
Cluster aware updating is based on the use of a new utility called the Cluster Aware Update Tool. The Cluster Aware Update Tool is automatically installed on all of the nodes in the cluster, but it is not active.
If you want to use the Cluster Aware Update Tool from outside of the cluster then you can install the tool by installing the Failover Clustering feature on any machine that is running Windows Server 2012 (you don’t actually have to create or join a cluster).
In order to perform cluster aware updating, the Cluster Aware Updating tool must run as a clustered server role. That way the update code can move from cluster node to cluster node as the updating process progresses.
How To Configure Cluster Aware Patching
To set up cluster aware updating, open the Server Manager and then choose the Cluster Aware Updating option from the Tools menu. When the Cluster Aware Updating tool opens, select your failover cluster from the Connect to a Failover Cluster drop down box, and then click Connect.
Once the connection to the cluster has been established, click on Configure Cluster Self Updating Options. This will cause Windows to launch the Configure Self Updating Options Wizard. Click Next to bypass the wizard’s Welcome screen and you should see a message telling you that the cluster isn’t configured with the Cluster Aware Updating cluster role.
You must now select the Add the CAU Clustered Role with Self Updating Mode Enabled to this Cluster check box. Click Next and you will be asked to set a self updating schedule. Microsoft recommends scheduling updates to occur at a time when there is the least possible demand on the
Click Next a couple more times and you will see a screen asking you if you wish to receive recommended updates the same way that you receive important updates. After making your decision, click Next and verify that the information displayed on the summary screen is correct. You can
complete the process by clicking Apply, followed by Close.
As you can see, it is relatively easy to implement cluster aware updating. After doing so, the process of keeping the cluster nodes up to date will be completely automated.