Windows

Remove Hidden Devices From Servers

One of the things I do after a successful P2V of a server is do some cleanup on the newly virtualized server. Cleaning up involves uninstalling any software , such as Broadcom  and Dell Open Manage software, that is no longer needed for the server. I also perform clean up of the any old or removed devices through the servers Device Manager that may still be lingering around. Removing these hidden or ghost devices can prevent or  fix any issues you may have with the vm.

To remove the hidden devices you will need to go to Device Manager and tell it to Show Hidden Devices, the devices will be displayed in a lighter shade than the installed devices. However, after performing many P2Vs , I found out that Device Manager doesn’t display all the devices that server may have had at one point.

Fig1.Hidden

Fig1.Hidden

 

 

 

 

 

 

 

Fig2.Hidden

Fig2.Hidden

 

 

 

 

 

 

 

 

 

 

 

One would think by selecting this view everything would magically appear based on the fact that we just told Device Manager to “show hidden devices”. Since we are dealing with Windows and something that would be logical isn’t, we can the modify the registry  to see these “hidden” devices . Once the registry value has been added ,reboot the server then you will be able to see all those hidden devices that were never uninstalled and you can properly uninstall them.

Add the following Registry Settings

Registry Settings
System Key: [HKEY_LOCAL_MACHINE\SYSTE\CurrentControlSet\Control\Session Manager\Environment]
Value Name: DEVMGR_SHOW_NONPRESENT_DEVICES
Data Type: REG_SZ (String Value)
Value Data: (1 = show all hidden devices)

Fig3. Hidden

Fig3. Hidden

 

After Rebooting the server go back into Device Manager < Select Show Hidden Devices again and you will now see the hidden devices shaded lighter.

Fig4. Hidden

Fig4. Hidden

 

 

 

 

 

 

 

 

 

 

 

 

 

 

To remove a device , right click on the device and select uninstall.

Fig5. Hidden

Fig5. Hidden

 

 

 

 

 

 

You may be prompted to confirm that you want to remove the device , select OK and the device will be removed.

Fig6. Hidden

Fig6. Hidden

 

 

 

 

 

 

 

You can remove hidden devices  on VMs or even physical servers  that has been running for a while as part of overall server maintenance plan to keep your servers running in tip top shape.

P2V’d VM fails to boot with incorrect disk controller

VMware converter tool is a really nice tool to convert physical server to virtual machines. It’s free and fairly simple to use. The majority of time that I have used it I have not had any issues but a recent issue where a P2V’d vm failed to boot inspired me to write this blog posting.

Back in 2010 when I was first implementing virtualization at my company, I ran into an issue with a P2V of aW2K8 server that wouldn’t boot up after completion. The VM would go into a continuous loop of blue screens and reboots. This was very annoying and made it difficult to determine the source of the problem.

The blue screen error gave the generic error that hardware on the server changed. DUH…I just P2V’d you from a Dell M610 blade so of course your hardware has changed, you silly server. The fun part was finding out what the server didn’t like about this change to cause it to blue screen.

The blue screen error code indicated something with the disk had been changed. With that lead I started looking at all of the VM settings, in particular the disk settings.  I noticed that the SCSI controller was set to Paravirtualized SCSI which was different from all of the other VMs I had P2V’d that day.  Their controllers were set as LSI Logic controllers.

I figured something must have become confused during the conversion of this blade, so I changed the SCSI controller on the VM to LSI Logic SAS within the vSphere client. The settings change worked, and the VM powered up normally.

Fast forward to August 2013 when I encountered a similar problem. This time I was converting a Dell 1855 blade running Windows Server 2003 R2 which also suffered from the infamous blue screen reboot loop. Having seen this issue before, I checked the SCSI controller setting and found that it had been converted as an IDE controller. Server operating systems are not compatible with IDE controllers so this VM was not going to boot unless I changed the adapter controller. When I attempted to change the disk type using the vSphere client, however, the option was not available.

ide1

A quick search on the internet led me to VMware KB article 1016192 (Converting a virtual IDE disk to a virtual SCSI disk).According to the article, if no controller is selected during the conversion process, the VM is created with an IDE controller for the VM’s system drive.

To fix the issue you have two options:

1.      Re-run the conversion, making sure you select a controller type. Do not leave it set to the default of auto-select.

2.      Manually change the adapter type inside the vmdk file.

I was under a time constraint and did not have time to re-run the P2V, so I opted to manually change the controller type. Using the instructions from the article, I was able to successfully change the controller.

To manually change the disk controller type I used the following steps:

Login to the host where the VM resides. If SSH is not enabled, you will need to enable that option before you can connect.

  1. Once logged onto your server navigate to the datastore path of the VM.

# cd /vmfs/volumes/<datastore_name>/<vm_name>/     ide8

  1. Using the vi editor, open the vmdk file of the VM with the following command.

vi nameofserverfile.vmdk

  1. Find the line that says :ddb.adapterType = “ide”ide6
  2. Change the adapter type to LSI Logic,

(type r over a letter then the replacement character until all the characters have been replaced with)
ddb.adapterType = “lsilogic” ide7

  1. Press the ESC key, then :wq to save the file.
  2. From vSphere Client:
    1. Click Edit Settings on the VM.
    2. Select the IDE virtual disk and remove it from the VM but DO NOT delete the disk.  ide2
  1. After the disk has been removed, you will need to go back into settings and re-add the disk.
    1. Click Add > Hard Disk > Use Existing Virtual Disk. ide3
    2. Navigate to the location of the disk you just removed and select to add it to the VM.
    3. Choose the same controller type from step 4. The SCSI ID should read SCSI 0:0. ide4 ide5

Change a Windows Server 2012 Product key?

For those that are just getting started with Windows Server 2012 you may have noticed that you don’t get the option to change the server’s product key on the activation page. This can be a problem if you need to switch to MAK or KMS keys and reactive the server. No worries, the solution is pretty simple.  You can use one of the following 2 options:

Option #1 –

1. Open a Command promtp and run the following :

slmgr.vbs /ipk  <put  your Product key here>

Option #2 – 

1.  On Windows server 2012  -( with your mouse point to the lower-right  hand corner of the screen)  click Search.

2.  Type Slui.exe 0x3

2012b

3.  When the activation screen pops up type in your new Product Key and Activate.

2012c

Expanding System Drive of a running W2K3 VM using EXTPART

At some point in time a VM will run low on disk space and you’ll need to add more space to it. For VMs running Windows server 2008 this isn’t an issue but anything older, well , you had to use tools such as DiskPart to get the job done. Let’s say the system drive was running low, for W2K3 servers, this was a little bit harder to expand because you needed to shutdown the VM. Depending on what you were running on that VM getting that shutdown approval could take an act of congress to pass. Expanding the system drive on a running W2K3 VM is not impossible though, for I have the solution that could solve those woes.

A few years ago when my company was just getting started on our virtualization infrastructure (VMware of course) I became familiar with a nice little tool that would expand drives on the fly for VM’s running W2K3 or older. This was pretty important considering over 80% of our servers at the time were still on W2K3. What’s this magical tool that can save us admins from having to shutdown a VM because the users hoarded all the space?

Dell’s EXTPART. I will tell you that there is no guarantees with this and use at your risk. Since I’m a daring person , well honestly, just don’t want to go asking to shutdown the server, I’ve used this tool. I have used it quite often too. I’m still alive, the servers survived, and half the time they didn’t even know they were low on space because I proactively fixed the problem.

Now, this isn’t the magical tool that fixes everything because it does have some caveats. For instance, you can only expand system drives on while running if the VM was not P2V’d. If the VM’s were built from scratch or from a template this tool works great but if your VM was P2V’d you need to reboot into safe mode to expand the system drive. Expanding non system drives while running is perfectly fine on P2V’d VMs using EXTPART.

To use this EXTPART do the following:

  1. To use EXTPART to expand a drive you will first need to download the tool and save it to a location on the VM. The tool can be found here
  2. In vSphere, go to the settings of your VM and increase the Hard disk that you want to expand. If it’s the system drive then most likely you will want Hard disk1.

    3. EXTPART_1

  3. Now go to your VM, open a Command Prompt and navigate to the directory where you saved the download.
  4. Once you are in the file directory of EXTPART, run the extpart.exe file
  5. You will be prompted to type in the Volume or Drive that you want to expand, type in C:
  6. After putting the volume it will display the details of the drive, such as volume and partition size, and ask how much you want to expand the drive. At this point enter in the size (MB) that you want to expand the drive by. This is the amt you are increasing the drive with not the total size of the drive.

    7. EXTPART_2

  7. Once complete press enter and the size of the increased volume will be displayed.
  8. If this fails with an error that “the disk is not accessible” then most likely the VM was P2V’d and you will need to reboot the VM into safe mode. Once booted into safe mode start at step 3 again.

So there you have it, an expanded drive on a running W2K3 VM. No time downtime and you look like a hero for expanding the drive. #WINWIN

Duplicate CMIDs and how they affect KMS hosts

The CMID ( Client Machine ID) is a unique ID that is used by a KMS host to activate the license on the machine. Typically each system has it’s own ID and there are no issues but with the use of tools such as cloning the CMID can be duplicated. When the KMS host treats each CMID as 1 client, so if you have 200 systems all containing the same CMID the KMS host has only activated 1 client machine. This results inaccurate licensing activation by your KMS host.

So what causes these duplicates you ask? Normally this is caused by some sort of cloning or imaging of a system that did not sysprep with the sysprep /generalize option, see Microsoft KB929829. I have also seen this happen when you deploy a vm from a vCenter template and you do not use the customization wizard to customize your vm.

vmcustomwizard

Once the CMID is set unfortunately Microsoft does not support changing the CMID, however it is possible to do so. Microsoft does not support changing the CMID and their recommended workaround is too re-image the system using the sysprep /generalize option because changing the CMID may cause the OS to become unstable. With a user’s system that is easier to re-image but let’s say you have several production servers that have duplicate ID’s?

Good luck with telling the application owners that you need to re-deploy their server to fix a duplicate CMID. I’m sure that will go over well. If you work in a company like mine,where that would not fly well, you have 2 options either you live with the duplicate CMIDs ( prevent future duplicates) or change it by resetting the CMID.

If you’re daring like me , I would look for the reset option and test it, to see what happens. Changing or resetting the CMID on a system is technically re-arming the licensing on the system. To re-arm the licenese you will need to logon to the system with the duplicate ID and run the following command:

cscript c:\windows\system32\slmgr.vbs /rearm

rearm

So there you go, you can change the CMID just keep in mind it’s not support by Microsoft if you do. The best thing to do is just avoid having duplicates by running the sysprep /generalize option when cloning systems. If you’re deploying vm’s just remember to run the customization wizard or have a precanned setting wizard all ready to go.