As a student I've not had much experience with the practice of setting up and managing networks. There's a question I've had for a while now, that my teacher wasn't able to explain to my satisfaction :

I used to be under the impression that server hardware had to have 99.99% uptime. I remember linux admins showing off that they had a box running for 3+ years. Imagine my surprise when I first found out that Windows Update for Windows Server 2008 offered a critical update to me every 2-3 weeks, and that most updates required a reboot.

When a VM asks for a reboot, usually it's no big deal as a second, redundant, VM can take over traffic until the first VM is back online. What I'm wondering though, is what to do with the Hyper-V server, the one that is hosting all those VM's?

My teacher told me that you can either reboot the server during off hours, which is reasonable for a small company, or not update that server at all. I can't even begin to imagine how unsecure a server like that must be after not having been updated for years, so I don't consider that as a valid option at all.

Isn't there a way to copy VM's to another physical Hyper-V server, and somehow have that VM not go offline to the end user?

I appreciate any advice and tips you guys can give me.


It is recommended to run the Hyper-V host in a Server Core install to reduce the attack surface and reduce the amount of updates needed.

With System Center Virtual Machine Manager (SCVMM) you could move the child partition to another host, as you suggest.

For further reading you can try TechNet or the 70-652 exam prep materials here.


Hyper-V supports clustering, which lets you move VMs to other nodes in case one fails or simply needs a reboot.


Windows prompts for the updates and to reboot mainly because a lot of the drivers are kernel mode ones. Linux runs lots of things in userland and can do updates without reboots. Even some kernel ones can go through without a reboot.

A lot of the machines that have 3 years uptime are custom kernels that have been hardened. I know I've a few linux boxes that have massive uptimes. One firewall box was running for around 3 years before the UPS is was connected to barfed. Its Kernel was compile with just about every feature removed other than those needed to do the firewall.

For the question, as Massimo said, Hyper-V supports clustering. So you setup two (or more) machines and then move the VMs across to other hosts. Install the updates and reboot the host. Then move them back.

Your biggest problem with always be with clustering of services. Exchange for example requires a lot of work to work across two boxes in a way that one can be rebooted without anyone noticing.

  • It should be noted that the FREE version of Hyper-V does NOT support clustering. But the version included as a role in server 2008/2008 R2 does. – Multiverse IT Sep 20 '09 at 22:07

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.