Kevin Remde's IT Pro Weblog
IT Pro Resources
IT Pro Evangelist Blogs
Today I thought I’d throw some good resources around Microsoft’s platform for virtualization: Hyper-V
“Hey Kevin.. how do I buy Hyper-V?”
You don’t. Hyper-V is free… well, more appropriately, it is just a role that you add to Windows Server 2008 or Windows Server 2008 R2.
Have you started using it yet? Getting good results? Give us your thoughts/comments/rants/raves in the comments here.
And in part 18, we’re going to show you how you can get a good idea on what moving a service to Windows Azure might cost you, compared to hosting it in your own datacenter.
If there is anything that Hyper-V really needs improving on it is the virtual networking. On a scale of 1 to 10 in maturity, i'd rate VMware networking at a 8 and Hyper-V at a 2.
Another thing that could use simplifying is Hyper-V form of VMware DRS. The configuration for DRS in VMware is a slider bar that is just managed with customizable rules for what runs where, what should run with something else, etc. Configuring SCOM to manage a Hyper-V cluster was way more complex and painful.
Anonymous - Could you elaborate on your comments and provide some examples? Without details, it's impossible to know the reasoning behind your statements.
Also, why would you be using SCOM to manage a Hyper-V cluster? Or did you mean SCVMM?
I will agree with anonymous to some extent. The current version of SCVMM isn't doing much to manage networking or DRS-like capabilities. Using something like SCOM and PRO-packs to monitor and then fire off tasks to address any load and re-balancing of virtualized resources isn't as simple as it could be. One thing I will tell you, though, is that SCVMM 2012 adds that capability. You'll manage logical networks, storage, loadbalancers, and other servers (for things like deployment of Hyper-V Server onto bare-metal hosts) and WSUS for deploying updates.. all from SCVMM 2012. (Beta is available now, by the way, and the product is supposed to ship before the end of 2011.)
I have seen a very DRS-like function demonstrated on SCVMM 2012 - where machines are shifted and re-balanced among cluster nodes. The example I saw was when an update needed to be applied to the hosts in a cluster. The process automatically Live-Migrated the machines around between the hosts as they were then patched and restarted.. in such a way that there were never any virtual machines stopped during the update process. And when it was done, it automatically re-balanced the workloads.
Hope that helps. I plan on discussing SCVMM 2012 in Part 27 of this series. Thanks for the comments!