Having been an early adopter of Hyper-V on Windows 2008, the improvements of Hyper-V in Windows 2008 R2 were definitely welcome with the Live Migration feature being the most useful in keeping virtual machines highly available.

Assuming you already know how to cluster your servers and understand CSV’s (Cluster Shared Volumes), configuring your server to be highly available is quite simple. You should carefully consider your host specs though because Live Migration doesn’t work with different CPU architectures. For instance, you can’t Live Migrate from a host using AMD processors to a host using Intel processors. Also, Live Migration is picky about the family of processors within the given architecture chosen, so make sure your processor specs are as similar as possible .There are ‘Processor Compatibility’ settings to allow you to migrate using different processor versions but if at all possible, try to purchase your hosts with the same specs to avoid not being able to use the Live Migration feature at all.

To use the Live Migration feature, you simply need to make your virtual machine highly available by choosing the ‘Configure a service or application’ under the ‘Services and Applications’ tab under your Cluster Management. The virtual machine has to be off in order to make it highly available. Once the virtual machine is configured and started, simply right-mouse click on the virtual machine and select “Live migrate virtual machine….” and select the cluster node. The amount of memory the virtual machine is configured to use determines the amount of time it takes to migrate it because in reality, that’s all that’s really being migrated.

Brian Pastre
Sr. System Administrator

If you’ve been around Windows servers long enough, I’m sure you’ve seen tasks you’ve grown accustom to in one version being changed to a location with no real reason in mind and have asked yourself many times “Why did you move that Microsoft?”. Print Management in Windows Server 2008 is no exception. Every time I open the Server Management console, I feel like I’m working on a Small Business Server and I find it takes more clicks to execute on the tasks that I need to do. For instance, it takes five clicks just to get to the place to install your printer drivers. Oh yea, I almost forgot - you have to install the role first!

So, if you are running your print server on an x64 platform and you have x86 clients that will be using your print server, make sure you deselect the x64 checkbox and select the x86 checkbox on the first screen after the welcome screen in the “Add Printer Driver” wizard. This will save you a lot of time.

In all fairness to Microsoft though, the ability to view the print jobs and view the printer administration web page after selecting the ‘Printers’ tab in the management console is very useful.

 - Brian Pastre, Premier IT, Inc.

One of the many roles server administrators are asked to provide now-a-days is high availability for those applications that run on the servers we support. Having a history of being associated with 24/7/365 and 8/5 shops, I have narrowed the goal of providing acceptable uptime down to three basic items; hardware, operating system, and applications.

Hardware:

Being in this industry for over 22 years, I have seen hardware make incredible advances in providing stability and I have seen servers in all kinds of environments. Even if your servers aren’t in a controlled HVAC environment, most will run very stable for years even in the most adverse environments. If you have physical access to your servers, you should make sure that your servers aren’t being choked. Keep the firmware and BIOS updated. Vendors have really made this easy compared to the early days where upgrading these could render your servers unusable.

Operating System:

Being primarily a Windows Server Administrator, Microsoft has made great advances in providing a stable operating system. I remember the days of NT 3.51 where one would have to reboot every week to fix memory leaks, etc.

Read the event logs, they tell you stuff! I’m amazed at how many server admins don’t read the event logs. I’m fortunate to have a monitoring system that alerts me for specific event ids, but I have been in environments where I had to read the event logs manually. Remember the old saying, “An ounce of prevention is worth a pound of cure”. Finding an event that a hard drive has failed in a RAID can prevent hours of downtime.

Patch your OS. As with any System Administrator, this is not one of my favorite tasks to perform and I think Microsoft releases too many but if you do them quarterly, you can still keep your SLA. The days of fighting through driver issues are pretty much gone but you still need to update them. I opt to use the manufacturer’s drivers over Microsoft’s.

Never waste a reboot! If you need to reboot a server for maintenance outside your normal window, apply patches before rebooting.

Applications:

Applications typically fall into three categories with respect to resource utilization; cpu intensive, memory intensive, and disk I/O intensive.  Knowing what your applications require before provisioning them into production can save you a lot of headaches down the road and help you achieve the SLA you're after.

 - Brian Pastre, Premier IT, Inc.