In this post I’ll talk you through how I have created a Hyper-V Failover Cluster on my single laptop.
To build a Failover Cluster you need a SAN. I use the Microsoft iSCSI target software that ships with Windows Storage Server. Being a Microsoft employee I have access to the bits, so I have installed it onto my laptop (so my physical laptop is going to be my iSCSI SAN. I have previously posted how to set all this up here. But I’ve re-posted it here to make your lives easier: Note that my offer of a time-bombed copy of the bits, for evaluation, is still valid.
Click to start, double click anywhere to play it in Full Screen and move your mouse over it to get the Player Controls to pop up.
On my demo laptop I have created two Windows Server 2008 Core virtual machines. Each VM is connected to the internal network switch that I created in Part 2 and is also connected to a Private Network Switch that I’m using for the Cluster Heartbeat. I’m using 192.168.0.x for my demo network and have uses 10.0.0.x on my heartbeat (I’ve also left IPv6 turned on).
I have added both Failover Clustering and Hyper-V using ocsetup to the Server Core VMs. I need to make sure you understand this bit – you cannot virtualise virtualisation. I can install the Hyper-V role into a VM but I can never start a virtual machine – if I do I will get a BSOD. Oh and install the RTM Hyper-V bits before you do this – the Release Candidate Hyper-V that shipped with the RTM of Windows Server 2008 do not let you do this (KB950050).
Once the two VMs are configured I create a cluster using the Failover Cluster Management tool (running from the Windows Server 2008 install on my laptop). Again, I have posted this before here. And I’ve re-posted the demos again to make your lives easier (these videos are old – still using RC code – but the process is fine).
Next I connect Hyper-V manager to one of the cluster nodes and configure a VM with the VHD and the configuration on the SAN drive. For my demos, I know I cannot ever start this machine, so I don’t bother installing anything onto it. Next I go into Failover Cluster Management and create a Highly Available Virtual Machine – which I can fail over from node to node (but don’t ever start it – BSOD on the cluster node). I have posted on this before here. And it’s here again (note that my comments about static MAC addresses have proved to be not true in the RTM of Hyper-V).
So now I can demo everything in Hyper-V. I have three Hyper-V machines to manage and I can show how to create highly available VMs.
In Part 4 I’ll explain how to get System Centre Virtual Machine Manager installed and working.