Security Analyst – Jeff Soehner CISSP, GXPN,GPYC,GMOB,GCIH

One of the things that can set apart a business is failover technology – lets’ face it stuff happens. Hard drives develop errors, power fails and your computers just stop. As a MSP I can wait to find out about it a few minutes after it does from my clients (who are not very happy because it failed) or I can find out with extensive monitoring that is constantly checking the hardware/software environment but I can’t really do anything about it unless we configure additional hardware/software to kick in when it happens. Until now those setups were large and expensive and involved clustering file systems, multiple servers and quite frankly would require dozens of man-hours to setup and maintain. Disaster recovery involved thousands of dollars per month to have on hand just in case you needed it.

Along comes Windows 2012 Hyper-V with features that help bring the cost of having the benefits of this technology to a small business for a few thousand dollars per year!

Enter Cluster aware updating – this new feature of Hyper-V allows any small to medium enterprise (SME) to implement two Hyper-V servers with enough memory to run all of the virtual servers they need to operate their business (Exchange, SharePoint, File server, SQL, etc.) Most systems come with redundant hardware (power supplies, hard drives, etc.) so a single point of failure is all but eliminated but what about updates and maintenance? If you want to keep everything running or even just want to test your recover options in the event of failure you need to fail over your live system with as little or no downtime right?

Let’s looks at an example below

Cluster Aware Updating initiates an update. It is determined that one of our Hyper-V machines needs an update so it is downloaded. CAU migrates the VMs off to another member of the cluster.

Next is a screen of the CAU progress as it begins to apply the update

After the update is applied CAU issues a restart of that member of the cluster.

Once the restart is complete for the patched Hyper-V machine all members are scanned again for patches

Migration back to preferred machines occurs once all Hyper-V machines have been updated and restarted.

All of this can happen automatically just as you schedule Autoupdates for all of your desktops. There is no need for an administrator to manage each patch and to manually apply them. You can still have little or no downtime and have complete fault tolerance for your SQL and Exchange databases which means no interruption for your business.

Now did I mention that it’s all free! Not one maintenance fee, not one software license or add-on cost associated with this solution – all you need are a few machines to run Hyper-V and your finished. Now I would like to see VMware come up with something that tops that!

I recently had a problem when I upgraded to Hyper-V 2012 with an older test machine (DC5750) that just wouldn’t work with the new kernel. I decided that it was time to create a new Cluster and dreaded the Import/Export that I was going to have to do. I thought this would be a great time to test the Replica feature. Not only will it allow you to replicate VMs onto another Hyper-V machine for disaster recovery but it will also allow you to replicate to another cluster!

Here I show you a snap of the Failover Cluster Manager with both of my clusters (MyCluster is the old cluster I am migrating from – you can see Hyperv8 is the only node and Cluster1 is the new one). You have to create a Replica Broker for Replications to occur on a cluster and I have now setup SW1 to replicate to the new cluster. Now I have all four of my VMs configured to replicate changes ever 5 minutes and can take my old cluster offline when I am ready. It will require me to start the new VMs on the new cluster but I can do this at any time.

Hyper-V Replica keeps a log of any changes to these VHD files and replicates those changes to the other VM. This results in a fairly quick time to recovery if any issues should ever happen to your main cluster. Initial replication can even be done offline to a disk and imported into the other site so bandwidth issues and large VMs are not a problem. You can even do all of this with a single NIC!

I am not happy that at the time of this writing they still do not have any management upgrades for Windows 7 (you will need to have a full version of Windows 2012 or Windows 8 to manage these new features) but I like the fact that the old Hyper-V management application will still allow you to work with your VMs.

Windows has hit this one out of the park and has challenged both Citrix and VMware to step up to the plate for most SMBs. I hope you find it as robust as I do and consider implementing your own CFS.