Hyper-V on the move

“Steady as she goes” could sum up the keynote speech at Microsoft’s MMS09 conference in Las Vegas. Microsoft has been doing great work on its management tools infrastructure, and now has a comprehensive range of capabilities for managing not just Windows servers but virtual machines too (and even out into the world of Linux). And on one pivotal morning, the speakers filled in some important holes and sprung some surprises on us.

First, the next major release of System Center will be called R2, a nomenclature favoured by Microsoft to signify new bits and pieces, but no fundamental redesign. Next they showed us live migration of a running Hyper-V virtual machine from one server to another, but I smelled a rat here – it looked too smooth with too little downtime, and this needs to be proved more convincingly in the future using real code. Moving live VMs was one of the features that Microsoft pulled out of the Hyper-V platform for its initial launch (for which it received a lot of criticism at the time), but it’s done and dusted now.

After that the company showed us a demo that raised the interest a notch higher – a live application move from one machine to another. This takes an application installed on one machine, together with all of its settings, dependencies and so forth, and wraps it up so that it can be moved onto another machine, installed there and run. The demo accomplished this in a few seconds (well it would, wouldn’t it?), but if this truly is a general-purpose tool its effects will be momentous. I think I’m stepping a bit too far forward here, but I’d imagine this is coming from the AppV application virtualisation toolset, which provides tools that wrap an app into a deployable stream. If so, the move feature is slightly less impressive, but still a welcome step forward, especially if it makes it more attractive for people to try AppV cocooning their apps today.

A later demo was pretty impressive as well – moving a VM from your “private Cloud” (that is, your own servers) to a third-party public Cloud provider on the internet. Of course, this happened in the blink of an eye, but I can’t believe this will be the case in the real world. Shoving a whole VM up your pipe to the internet isn’t going to happen in a snap of anyone’s fingers, unless you have a really huge pipe. I certainly can’t see this working over your average ADSL Max line.

Overall, the news so far was good: solid work, delivered in a timely fashion, but nothing too surprising. This is, after all, a management tools summit and surprises aren’t welcomed by the sysadmins who attend.

Best Practice Analysers

Kudos to Microsoft for continuing to work on its Best Practice Analysers, the tools you run on your server to see whether various Microsoft server technologies are actually set up properly. My Exchange Server 2003’s event log was filling up with red errors that related to the rendering of IMAP messages, but before deciding to dive into the server configuration and give it a thorough spring clean, I decided to try the latest version of BPA. It had a good poke around the box and reported back a pile of changes it thought would help. I followed up these suggestions and then decided to use some of its benchmarking tools to see just how quick my hard disks are. Again, lots of useful stuff was reported. Microsoft really should better integrate this tool into the main products, since it covers a lot of things that you should be looking at every few months. Think of it as an MOT for your services.