Ghost Solution Suite

I have a lab of 31 identical OptiPlex SX280's running XP SP2. I have had an enormous amount of problems with GSS in the last couple of weeks that seemed to start kind of randomly. In attempt to hopefully correct some of the problems, I upgraded to GSS 2.0, but that doesn't seem to have helped. I tried completely uninstalling ghost client on all of the machines and ghost console off of the server and reinstalling, but I still have problems. My current problem is this:

It will only recognize one of the machines at a time. We are currently using the temporary unlimited license while we wait for our 40 license permanent code. It will allow me to install the ghost console client on all of the machines, but only one will actually show up in Ghost Console. More specifically, the first machine that I install the client on shows up in ghost console. However, if I try to send a task to it, the task actually executes on the LAST machine that I installed the client on, even though Ghost Console still shows the name of the first one. If I uninstall the client and delete the machine from Ghost Console, then another one of the machines that has the client installed will show up, but still only one. The machines do all have unique MAC addresses, IP addresses, and host names. If anyone has any help on what could be going on, it would be greatly appreciated.

Also, the last time I imaged (before I uninstalled and reinstalled, it was still talking to several of the machines,) Ghost successfully cloned all of the machines, but then failed to add them back to the domain (it did set the names, DNS servers, etc., though.)

I had never had this much trouble out of ghost before a couple of weeks ago. If anyone knows what might have happened, again, please help me out here.

Nevermind, I figured it out. It was the TAP-Win32 adapters that were added to the machines as of the last image, which would explain why it broke then. Apparently, they were all using a default MAC address when they were not being used for anything and the clients were sending that to Ghost as their MAC address, thus Ghost thought they were all the same machine and, when a new machine installed the client, Ghost just updated an existing computer record with the new machine's info, which is why the last machine was replacing the machine in Ghost Console. After uninstalling the TAP adapters, Ghost works fine again. Looks like I'll have to figure out a way to script the installation and uninstallation of TAP adapters, but that would be helpful for our purposes anyway.

Thanks for the heads-up on that; I do try to filter out the MAC addresses of things that aren't real adapters, but it seems that is managing to get through.

Tunneling adapters like this have a specially assigned type code in Windows, and we filter it out. This software adapter isn't using it, and is claiming to be a real Ethernet adapter, with a MAC address of the form 00-FF-xx-xx-xx-xx, where the x's are based on a random GUID that Windows generates for adapter installs.

I'm a little nervous about filtering out this type of MAC address, because it's scattered throughout the range of valid, real assignable MAC addresses. Every other kind of address I've been forced to filter out has either used a properly assigned IEEE OUI for its addresses, or only occupied a tiny part of the address range, and the only other thing I could match is a string that can be localised.

However, I've checked the IEEE registry and it seems that 00-FF-xx hasn't been assigned to any company as an OUI, which means that although this could cause us trouble in future we should be OK for now (and in future more machines should have properly unique SMBIOS UUIDs so we don't need to rely on MAC addresses).

I'll add this pattern to the existing filter list in the clients; this isn't as easy to update as the server, since the clients are embedded in your images, but if you have trouble with the scripting solution we should be able to work out something if you get in touch with me directly at nigel dot bree at gmail dot com.