After performing a physical to virtual conversion one of the tasks you have to do is to set the IP address of the new virtual NIC. When you do this you then get a warning message saying the IP is already assigned to the physical NIC, Windows 2008 is better than 2003 as it at leasts removes it from the old NIC for you.

All the same this seems a bit illogical, that physical NIC is no longer connected to the server so why should it keep the details? Also it can cause confusion for some services as the details are still stored in the registry even though the NIC isn't active. Should you have a dig around there's no trace of it in the Device Manager though, and hacking bits out of the registry isn't a great idea.

The process for removing the old NICs is actually quite simple and I found a good explanation of how to do it here:

http://www.tacktech.com/display.cfm?ttid=419

I've been migrating a client to vSphere 4 this weekend and have encountered this problem with all their servers, and its not the first time either so I thought I should share the solution.

This is most common with Dell servers and their so called "OEM" licensing, and its another one of their sly cost-saving measures that you dont notice until they really cause you a hassle. To elaborate a bit more, you  may or may not be aware that OEM stands for Original Equipment Manufacturer, and such a license is supposed to only be used once on the hardware it is supplied with. HP are more honest about their licensing and offer two types of OEM license, Standard and "BIOS Locked", which sometimes appears as "Restricted OEM" if you run a system information tool such as Belarc. The difference is quite major, standard OEM will require re-activation after a significant hardware change whilst BIOS Locked will not re-activate after a change. You can guess which version is cheaper, hence why Dell use it, but its hard to tell beforehand as they do not tag it any differently.


Its always a good idea to use a tool such as Belarc to get a full system report before you start a conversion, here's part of my one:


Dastardly Dell fail to mention that this is actually an OEM-Restricted license....


Since converting a server to a virtual machine represents a significant change of hardware it will initiate the re-activation process, with standard OEM you will just get the popup reminder after you have logged in. BIOS locked OEM however will not let you get past the logon prompt, even in Safe Mode, you have to complete the re-activation process or shut the system down. Seems simple enough until you go through the process, probably having to use the "Telephone" option as you can't configure the network properties beforehand, and after typing all the numbers you will end up with a Microsoft operator telling you to contact Dell (or the original hardware supplier).

You've probably already spent all the time on the phone to Microsoft and Dell getting nowhere, but if you haven't yet then don't bother. The only way you will get it to activate at this stage is to obtain a new OEM product key and I haven't found a way to convince them yet. There is a not too painful solution though, but it might cost you a new license:

You need to obtain a license key for the version of Windows that you are running that you know will activate and the media to go with it. It doesnt matter if your VM is Windows 2003 R1 and you can only get an R2 license and media, this process will still work and you'll get an R2 update as a bonus. Try and make sure that the media is the same service pack level as the VM though otherwise you may get a lot of errors until you can reapply the service packs later.

Boot your VM off the Windows CD you have obtained, if you are using ESX then you will probably have to go into the virtual BIOS and change the boot order first, and the install process will start. Select the option to install Windows and it will run a scan for existing installations on the hard disks, if at this point it says it can't find any hard disks then you need to check what the virtual SCSI controller is. This is only likely to occur with ESX as HyperV and Xen both use IDE, if the controller is "BusLogic" then try changing it to "LSI Logic Parallel" and you should find the installer recognises it next time. 

The Windows Setup routine should find your existing installation so select the option to repair it, then let it run through the re-install process. About halfway through it will prompt you for a product key, enter the new one you have obtained and it should accept it, if not then you haven't got the right media to match the key type - OEM keys need OEM media, Volume License needs VL media and retail needs full retail media. 

There is another issue which sometimes pops up if your server has .Net 2 installed, you may be prompted for the CD2 media as it wants an "install.exe" file, but this isn't anywhere on the Windows Server media. To get round it you have to download the dotnet2 redistributable and use a program such as 7zip to unpack it (you'll see the install.exe in there!) and then another program such as ISOrecorder to turn it into an ISO. Mount your new ISO on the VM and the installer will happily continue, I've also heard the same thing can occur with IE 7 or 8.

Once the installer has completed the server will reboot and this time you should be able to login without any problems. Before you do  anything else though I advise running Windows Update as the repair process will have effectively uninstalled every update released after the media date.

We've just acquired a pair of HP DL380 G6s for our vSphere 4 test lab, they have the new Intel Nehalem Xeon CPUs and chipset with all the latest virtualisation features. Plenty more to come on all that but first of all here's something you won't find in the documentation. Having just installed the ESXi 4.0 hypervisor on the servers the classic question came up from one of the junior technicians - "what's the difference between ESX and ESXi?". Someone else then answered that ESXi doesn't have local console access like ESX does, so all management has to be done from the vSphere client. Although this answer would probably score you points in the VMware exam its not exactly true, as I was able to prove.... This is the same procedure which worked in ESXi 3.5, it still works in 4.0 too. Although VMware do provide a remote CLI appliance that gives you much of the scripting support you'd get with ESX sometimes its much easier to use the console, so once you've got your ESXi4 server up and running at the yellow and black screen do this:
1) Press Alt-F1 2) Type "unsupported" - note you wont see this echoed on the screen, but after you press return you will be prompted for the root password. 3) Enter the password 4) Marvel at "tech support mode"

Whilst you're in there why not give yourself remote SSH access too:
1) type "vi /etc/inetd.conf" 2) remove the # before the SSH line, save and close 3) enter "services.sh restart"
The VI text editor is a good example of how ESXi has been stripped down, its the old school *nix text editor and you will probably have to Google the instructions for it if you haven't used it for a while, but forget about something fancier like Nano. Like its says on the warning page though, this is "unsupported mode", so if you break it don't expect to get any support, you'll just have to re-install. ESX Server was has always been in effect a glorified Linux server, with all the admin and configuration options you would expect from a Linux server. The problem VMware had was that it was getting rather bloated, a standard ESX 3.0 install would use near enough 2GB of disk space and took a good few minutes to boot. In response to this VMware developed ESXi, a lean, mean hypervisor with minimal disk and memory footprint which could then be "embedded" on a server. This fitted in with VMware's vision of the future of virtualisation, where server boxes just become plug and play vSphere resources. It also happened to reduce the functionality enough to not annoy all those people who had paid for ESX when they started to give ESXi away for free :) Reading the VMware documentation you get the impression they rebuilt the ESXi hypervisor from scratch as a dedicated virtualisation OS. In fact it looks more like they went through their last Linux build and simply removed all the stuff which wasn't essential. As a result there is still a Linux based OS under there, and a console you can access, but a lot of the packages you might expect as standard won't be available.

Although proper SSL certificates from providers such as Digicert and Verisign have become much cheaper nowadays there are still many occasions when you might prefer to just use a self signed certificate. One of the most common occurences of this is with SBS server, to provide SSL secured web services such as OWA and RPC over HTTP. However the certificate warning you get everytime you open the web site is annoying and maybe confusing for less experienced users. In older versions of IE you could just install the certificate after viewing it but with the improved security in IE7 and IE8 this no longer works. Whilst putting the URL in your "Trusted Sites" list will let you view and import the certificate you will find it is still not properly trusted. The solution is to provide your users with the correct certificate file to install, but its not obvious where you get this from....

One complaint our helpdesk hears quite often from clients is that "our Internet access seems really slow today", which is a particularly vague and subjective problem. Once the technician has established that this condition is affecting all users and isnt just a temporary period of high demand what can he do to establish the cause of the problem?

A few months ago we took over the IT support for an insurance firm in the City, but with the contract we inherited a problem the previous IT company couldn't resolve. Part of the firm's business still ran on a ten year SCO Unix server which was barely clinging to life and hadn't been backed up for three years. They still had a support contract with the original suppliers of the SCO system but they had quoted over £10k to migrate their data to the current platform. The system was running on borrowed time and every reboot could have been its last so we had to come up with a solution.

Petri have just published my latest guide which shows you how to setup authenticated and anonymous SMTP relay services on an Exchange 2007 server.

 You can view the article here

This is a subject I haven't looked at for some time, maybe surprisingly its something we don't get asked about all that often. A typical scenario might be where a company has an Exchange Calendar Public Folder containing details of company events, important dates and so on. They want their mobile users to be able to view this calendar on their Blackberrys or Windows Mobile phones so how do they go about it?

LES circuits have become increasingly common in recent years for providing high speed WAN links between two sites, usually either between two company offices or between a client site and an ISP PoP (Point of Presence), where it is used to provide Internet connectivity. The limiting factor is usually distance to the nearest PoP as the LES circuit cannot cover more than about 15km so they tend to be limited to city environments. It is based on fibre optic technology but the end connections are presented as standard RJ45 network ports, standard speeds are 10 or 100Mbps, so for practical purposes it is the same as a long CAT5 cable.

There are however a couple of things to look out for when setting up a LES circuit, and there is little documentation generally available:

I'm not a fan of catchall mailboxes, they are a legacy of POP mail systems and getting Exchange to do it has always been a bodge. However the need arose recently with a client we migrated to SBS2008 so I decided to see what the options are for doing it in Exchange 2007.