Archive for the 'IBM Stuff' Category

Hyper-V host blank black screen

I recently had a problem with a couple of IBM Blades that I was trying to deploy as Hyper-V hosts.  I employ the use of a replay volume from our Compellent storage to create an image for my blades.  Basically, I install one, sysprep it, and then copy the volume mount the copy as the boot volume for each of my blades.  The most recent hosts that I attempted to use this technique with, would boot, but once they made it into Windows, the screen would go black and there would be no way to interact with the machine other than turning it off.

This happened on two blades in two chassis so I assumed that it must be the image.  I made a new image, and it worked just fine, until I installed the Hyper-V role.  Once I installed the Hyper-V role, the machine exhibited the same behavior as above. 

With a short amount of searching, I came across this:

I have actually seen this mentioned before for something else, I just don’t recall what, but basically the problem was that the two processors in each blade were different revisions.  The simple solution we employed was to swap a processor from each blade to have the two processors match.  Now all is well in the land of Hyper-V.  At least for the moment…

An online pack already exists…

Situation:  I have a 32bit file server running Windows Server 2003.  I have a new 2008 R2 (x64) file server running on Hyper-V.  Did I mention that the R2 server is on Clustered hosts?

This is all simple.  Just drop the LUNs from the file server, give them to the cluster and assign them as pass through disk to the guest OS.  Simple…

Except, I kept getting the following errors when I try to do the import:

An online pack already exists.

The operation failed, because an online pack object already exists.

The provider encountered an error while merging two packs.

I mean really…  what does all that mean.  I couldn’t find anything pertinent in my searches, I tried working in Diskpart, and got some of the same errors.

After lots of searches, I couldn’t find anything to help.  So I opened a ticket with Microsoft.  I got a very helpful person who told me that it basically sounds like a driver issue, but he couldn’t say for certain.  He couldn’t find much on it either.

So on a whim, I moved my test LUN from my production file server to another 2003 server.  That worked just fine, so I then moved it from there to the 2008 R2 machine that I had been working with.  That worked just fine.  Weird…

So I checked the driver for the production file server:


and compared that to the other 2003 server:


As you can tell, the culprit seems to be an outdated driver on the current file server.  Seems I don’t keep things as up to date as I should.

How to remove a Persistent Reservation on a DS4700

We use a couple of IBM DS4700 storage controllers at my work, and I had one configured with a Windows 2008 Hyper-V cluster.  The cluster was for testing (I am building a new one on R2 for production) and when I went to remove everything that had been used for testing, I came up against a problem:

Error 219 – The operation cannot complete because the logical drive has a persistent reservation placed on it. Please release the reservation at the host and then retry the operation.


Well, maybe I should have told the cluster to go away before I started deleting the storage, but I didn’t so, that seemed to be a bit of a problem.  I did some google searching, and that was very frustrating, because although I did finally find a reference to it, the reference seemed to allude to the necessity of calling IBM support to correct the problem.  If you found this post, and you are having the problem, please be comforted by the fact, that you DON’T have to call for support.  I am running Storage Manager 10 and the latest firmware, and if you go to the Advanced –> Maintenance –> Persistent Reservations… you get this:


and if you click in the “View associated registrations” box it will show you which host/adaptor holds the reservation.  But the really cool part is you can highlight the LUN you want to remove the reservation from and click “Clear”.

IBM NetBAY Virtual Console Software and C2T

If you use IBM hardware, and you happen to have a bunch of xSeries servers, and some of them happen to be x335 or x330’s with the C2T cable management…

To be able to quickly and reliably switch between the servers in a stack, you need to create a new macro to count for you.  Let’s assume you want to be able to switch to the 3rd server in the stack.  From your console you have to enter numlock, numlock, 0, 3, enter.  Easy enough, but depending on network conditions you may have to enter that a number of times before it actually takes. (Trust me this is frustrating.)  But you can create a macro that will do it for you. 

Simply go to – Macros: Configure…:  Create.  From there type in what you want to call it (I named mine “C2T 03” for this example). 

In the “Keystrokes” box enter the following:  <Num Lock><Num Lock>03<Enter> and  then hit ok.  This isn’t the end though, because when you create a new macro it doesn’t just pop into your list.  You have to edit the group or create a new group (which is what I recommend) so that you can add the new macro to your dropdown list.  If you have questions about this you can e-mail me and I will try to help you out.