imageAs cool as writing and working with this series has been.  I love to talk about technology, especially technology that can help save you time and money.  When you look at the System Center products and a tool like PowerShell how can you not enjoy saving time and money.  However one of the key things I have learned throughout the process of working with all of these technologies is that  it all boils down to the basics.  In the world of the IT Pro those basics are all centered around the network.  From DNS, to Active Directory, to fail over clustering, to email….etc.  None of it works without  a healthy and properly functioning network.  Aside from the occasional I unplugged the cable, see one of my early forays into writing: Field Notes: The Doctor Is In.  you really need to dot all of your i’s and cross your t’s when it comes to networking.

For me one of the facts about virtualization that makes me smile is that you still have manage and worry about networking to the servers.  This environment can get tricky when you start working with limited physical resources.  For example take a look at ,y current event environment.  I have two laptops, and have to be able to showcase Live Migration.  For me this meant I would be leveraging my virtual environment to host all of core services, AD, DNS, Storage (using Microsoft iSCSI Software Target 3.3), SCVMM, and a SQL server.    I also had enabled failover clustering  on my two laptops.    I was able to get all the machine domains joined and configured.  I even configured my DC (as well as my other virtual servers with a delay start) to come online automatically after every boot, that way I can log on to domain when I restart my laptops. 

Now I was ready to setup my clusters and get them all connected and try out my live migration.   Before I go on with all the fun I had, I have to give mad props to Aidan Finn, for his Hyper-V clustering guide which you can find imagehere:  How to Build a Hyper-V Cluster Using the Microsoft iSCSI Software Target v3.3 Aidan’s guide is a great place to start if you want to kick the tires with live migration.  I learned quickly that there is a reason why we recommend 3 physical NIC’s for clustering.  As I was going through the steps in Aidan’s doc this is when I was reminded of something my good buddy Keith mentioned to me, Core Networking is a Skill.  It is not just about making sure you can do simple pings but also making sure you have proper name resolution and routing done properly.  You also want to make sure you plan and diagram your network before you build it, that way you can see it and something to refer back to.  A basic diagram of my network is to the right, Aidan’s doc is much nicer.    Lastly and most importantly you have to have the skill to interpret error messages that sometimes are a result of bad networking.  Here are some of the my favorite examples that I saw as I was having fun building this environment:

    • "RPC server is unavailable."
    • “Migration Failed”
    • Not seeing Servers Cluster to install the SCVMM agent on.
    • "The program cannot open the required dialog box because it cannot determine whether the computer name “Your Computer Name” is joined to a domain."

The last one is my favorite of the bunch.  I got that error message from a domain joined system (HVHOST1), logged on as the domain admin.  My first thought was what what, there is no way you are joined to the domain.  I may used your more colorful language but you get the point.  As it turned out my cluster name was resolving to wrong IP address and my cluster I made a typo for the access point IP.  sigh.  But when you look at that error, trust me IP configuration  is not your first thought .  But this goes to prove my point make your network solid and name resolution great, then you have a solid foundation  to build other technologies on.

One of the last things very unique to my network is the setup and since I do not have the recommend 3 nics how could I get live migration to work.  Again this was all a networking issue and understand how clustering leverages the network and was key to why my live migrations were failing in my little virtual network.   Well as it turns it the clustering service automatically chooses the preferred network and you can see which one is being used with this PowerShell command:

Get-ClusterNetwork | ft Name, Metric, AutoMetric, Role

The one with the lowest metric is the one that will be used.  You can control this either with PowerShell or via the GUI.  For my I had to basically configure my networking to use the physical network than viola success!  Again it starts with a solid network, and knowledge how the servers and services are going to leverage your network. 

There are some great resources here to help with your clustering and networking:

As I mentioned in part 1 we are going to be posting the series across 4 blogs: Brian Lewis, John Weston, Kevin Remde and myself. Let us know what you think of the posts and if you think of topic let us know!

If you happen to miss a part and want to get caught up. You can find all the parts of the series here: The Cloud on Your Terms: 30 Days about the Cloud