One of my new favorite features of Virtual Machine Manager 2012 is the ability to cluster the Virtual Machine Manager Server and Libray so a single host failure wont impact the ability for VMM to function , this also make improves of the uptime at patch windows , as we can move the VMM to the passive node patch and rinse repeat

This short introduction will go though the steps to setup clustering on SQL and VMM but due to limited hardware resources it will be configured a non clustered Microsoft ISCSI Target.

This is a walkthrough there are room for improvement (multiple network cards , proper SQL design , SPOF) ,  so its meant as a short demo on how easy the cluster enablement of Virtual Machine Manager is

VMM require Windows Failover Clustering with a cluster running before installation of Virtual Machine Manager.

First we need to create the ISCSI targets for the 2 VMM nodes and the 2 VMM SQL Server Nodes , again this is testing so the ISCSI goes on the same network as all other traffic not something that will do for production or large scale testing, but this is a playground so we can cut some corners.

And we need some disks for the clusters , we just adding a quorum and a data disk on both the SQL and VMM cluster

And we need to present the disks the the iscsi targets.

And then we are ready to form the Windows Clusters and get ready for SQL and Virtual Machine Manager Installation

On each of the 4 machines we need to configure iscsi and connect the drives before starting the cluster installation.

And create a Qourum and Data drive for the SQL installation

And verify on the second node that the luns are visible.

And enable Windows Failover Clustering Feature on the VMM and SQL nodes.

Add both nodes for creating the cluster

And let the cluster validation wizard do the work for you to ensure that the cluster can be formed

And we are ready to form the cluster , we have one networking warning due to no redundancy of the network setup again for testing we can live with it.

And we need a name for the soon to be formed cluster

And the wizard is forming the cluster for us

And on the SQL node we needs to start the SQL Failover Cluster Installation

We need SQL Database Engine and Management Tools

And we need a name for the clustered SQL Server

Default is that we create a new cluster and let the install add the resources needed

And we need a cluster drive for install

And a ip address for the SQL part of the cluster

And we need to ensure that we use Latin1_General_CI_AS for Virtual Machine Manager database

And the we are waiting for the install to proceed

And Sucess

And on the 2nd node we need to start the installer to ensure that we have a 2nd node to failover to if the hardware breaks down

And we need to ensure that we install to the right instance

And waiting

And Scuess we now have a running SQL cluster

On the VMM nodes we need to setup the shared storage and start the failover cluster wizard as we did on the SQL servers

And we need a name for the base part of the cluster

We need to install WAIK on both nodes before we start with the VMM install

VMM detects that its running on a cluster and ask if we want to install into a cluster to ensure HA incase of hardware failure

Default install location for the application

And we need to point to our newly created SQL Cluster for placement of the database

And we need a name for the VMM Cluster itself

And for VMM HA a domain account and AD bound encryption key is a requirement

Library shares needs to be created manually after VMM Instal

And we are installing.

Success

On the second node we need to run the VMM installer , this will detect that VMM is already installed as HA and can then add the 2nd node to the cluster

The Wizard picks up the SQL information so no interaction is needed (exept for next)

We need to enter the password for the service account

And the installer is doing its job

And Success time to test

When starting the VMM Client we need to use SCDVMMCLUSTER and not the local nodes in the cluster

And we can see that we have connection to the cluster and we now have a VMM infrastructure with no single point of failure