Archive for August, 2011

Silently activate Windows 7 in SCCM OSD TS

August 25, 2011

When you want to roll out Windows 7 to a couple of workstations, but do not have enough of them to justify a KMS server, you are forced to use a MAK key.
No problem, but activating the key after the installation requires Admin rights, so i cannot be done by the user itself.
If only we could do this silently as part of our SCCM OSD TS…
Fortunately this guy made a nice thingy that prevents the “Congratulations, Windows has now been activated!!” screen from f-ing up my silent TS.
Nice!

VAMT 2.0: Remaining Activation Count

August 25, 2011

What to do if you entered your product key correctly in VAMT 2.0, but when pressing “Refresh Product Key Data Online” you get ?

Two tips:
1. Set your proxy (if any) correctly
2. Set your locale to English – United States

(Control Panel, Region and Language, on “Format” select “English – United States”
Then in VAMT click “Refresh Product Key Data Online” and voila, it works.

I wouldn’t know why there would be a difference between the Dutch and the English Locale for just displaying a stupid number like 5000, but hey, appearantly this is nescessary…

Thanks to Les52

SCCM: No Boot Action for Device (number) found

August 25, 2011

I had a computer on which the OS was damaged so it would not boot, and using PXE boot it would just get the “abortpxe.com” so it could be reimaged.

How do we know what collection it is in, so that we can find out why is doesn’t have any PXE advertisements?
– There is no label on the computer with a computername
– The OS won’t boot, so we cannot see the computername within windows.

That leaves us with only the MAC Address and the GUID to search with.

When you do a search in the SCCM Console in a collection , you can not search on either of these values.
So we have to use a query or a Report for this.
You can do this by using a standard report, within the folder Network, called MAC – Computers for a specific MAC Address:

But only if the system is known within SCCM.
So what if this does not show any results?

Then the system is unknown within SCCM, and therefore automatically is a member of the collection “Unknown Computers”
So any advertisements targeted to that collection should also apply to this computer.

But what if it still gets the “abortpxe.com” stuff?
Then it’s time to get into PXE.
Open the file smspxe.log which is located on the sccm server in \SMS_CCM\Logs
Search in there for the mac address and for any messages.
In this case, it showed that sccm thought this computer was a known computer, with no active advertisements.
These are the messages in the log:

Executing LookupDevice..
CDatabaseProxy :: LookupDevice succeeded:
Device found in the database. MacCount=1 GuidCount=0
Executing GetBootAction
No Boot Action for Device (5292) found]
ProcessDatabaseReply: No Advertisement found in Db for device]

So, the device was foun din the database, on the basis of its MAC Address.
Strange, because I could not find the computer anywhere in the SCCM Database.
Now there is some caching of the SCCM database in the PXE part, so maybe it was once a known computer, but it was deleted recently.
So, a restart of the PXE service (or rather the Windows deployment Services – WDS ) on the sccm server should fix this.

This didn’t help either, still getting the abortpxe.com
Well, if SCCM thinks it doesn’t know the computer, but PXE thinks it does, let’s tell either one of them something else.
So I added the computer to SCCM manually, by creating a new Computer Association.
Then added the “new computer” to a collection with the correct advertisements, and voila, we have working PXE boot again.

But the question remains: how can PXE/SCCM “find” the computer in the database, while I cannot find it anywhere?

SCCM OOB console fails “GetAMTPowerState fail with result:0x80070005

August 23, 2011

So i implemented OOB management in a SCCM 2007 R2 SP2 environment. I had set up the whole PKI infrastructure, opened up the nescessary network ports, and everything worked fine.
Up untill now.
When i try to connect to a workstation using the OOB management Console, it would hang on “Connecting…” and then go to Disconnected.
But when i rightclick a computer, select OOB Management -> Power Control, i AM able to turn the computer remotely on and off.
So there is communication, but the OOB Mgmt Console can’t establish a connection to the workstation.
The logfile oobconsole.log located in C:\Program Files\Microsoft Configuration Manager Console\AdminUI\AdminUILog shows the error
GetAMTPowerState fail with result:0x80070005
Now, i’ve searched a LOT on the internet for a solution, but haven’t found one.
I checked the Kerberos tokensize, which was well within bounds, also i checked the PKI CA if it was issuing the SSL certificates to the clients properly, and ran through these excellent SCCM OOB Troubleshooting posts (part 1 and part two
Still, same error.
So i ended up manually going into the AMT in the client machines, and selecting Unprovision, (Partial) and waiting for the next provisioning cycle to pick them back up.
Because this takes 24 hours normally, i changed this to once every hour, to speed things up a bit
After that, i could connect to them without problems again.

SCCM: BITS download failing / hangs

August 5, 2011

Classic problem today: when downloading a program from the “Run Advertised Programs” item in Control Panel on a ConfigMgr client, the download of the software “hangs” .
When cancelled and restarted, it hangs again, although at a different percentage of the download.
Following this post quickly turned me to the ApplicationHost.config file, where i had to allow another filetype to be downloaded through IIS.
This time is was the “.config” file-extension and the “.resources” file extensions that caused the 404 errors in the IIS-log.
These files were appearantly present in the Intel AMT / IME drivers/application which i had packaged.
After modifying the ApplicationHost.config (on all DP’s!)and doing an IISRESET, the BITS download worked perfectly.

SCCM PXE boot fails: abortpxe.com

August 5, 2011

Just a short post to help out some people who want to PXE boot a machine to do a SCCM OSD, but get this error.
Tip:

Doublecheck which SCCM Collections your machine is a member of, AT THIS POINT.

Let’s say you made an OSD task advertisement to the collection “All unknown computers”
You have a new computer, boot it up in PXE, and yes, you get the Advertisement for the OSD task.
The task runs, the computer gets windows installed, other software, and the sccm client.

Then, you want to run the same task sequence again on this computer, through PXE.
But, when you boot from PXE, you get the Abortpxe.com, and it won’t boot from PXE?
Why?
Well, because after you ran the previous OSD and the SCCM client was installed, your computer is no longer “unknown”.
It is now in the SCCM database,
and therefore no longer an “Unknown computer”,
and therefore it is no longer member of the collection “All unknwon computers”,
and therefore it will no longer get the advertisements which are targeted to that collection.

Actually makes sense doesn’t it? 🙂

So, find out to which collection the system now belongs, (at least it is a member of “All systems” now, but you wouldn’t want to advertise any OSD tasks to that) and create a new Advertisement for an OSD task to that collection.
Now reboot and try again. Voila, there is your PXE boot again

Event-ID 25: Volsnap. The shadow copies of volume x were deleted because the shadow copy storage could not grow in time

August 5, 2011

I had a Windows Server 2008 R2 running Symantec BackupExec 2010 R2, doing Backup-to-disk-to-tape jobs daily, running fine for about 6 months. Then, for some reason the backups started failing on a part of the job.
Specifically, it failed on the part where it had to backup data from a local drive on the server (G:\) to the Backup-to-disk folder which was also on the same G:\ drive.
This had worked fine for quite a while, but somehow it started failing now, with the following error in the Windows Event Log:

Event-ID 25: Volsnap. The shadow copies of volume G: were deleted because the shadow copy storage could not grow in time. Consider reducing the IO Load on the system or choose a shadow copy storage volume that is not being shadow copied.

This would appear to be pointing to the drive not being able to handle all the IO traffice of reading from it, and writing to it at the same time. But it had worked for a couple of months before, so why would it suddenly not be able to handle this load then?
First i update BackupExec to the latest patchlevel. No difference.
Then i noticed there was not a lot of free space on the drive anymore, out of the 4 TB there was only 200 GB free space.
Maybe this wasn’t enough for BackupExec to be able to work with, so i cleaned it up to about 600 GB free space, and set the Minimum free space option for the B2D folder to 400 GB, so it would start removing old B2D files sooner.
After that, tried the backupjob again.
Still failed, with the same error. 😦
Then i did some reading on vssadmin.exe, and what can be done with it.
I used vssadmin to change the location of the shadowcopy storage.
This is where the vss service stores it’s temporary files while making a VSS backup of the drive.
The default location is on the drive itself.
So i changed this to another drive on the server which had about 200 gb of free space, and tested the backup again.
This time it worked! No more vss errors.
This change distributed the load of the VSS process amongst the 2 drives (read sourcefiles from G:, write shadowcopy of them to E:\, then write backup of E:\ to G:\B2D again), and the backup could successfully complete.

Intel IME / IME / HECI Windows XP driver issue on HP 8200 Elite SFF pc

August 2, 2011

When adding the Intel IME driver for Windows XP on a HP Compaq 8200 Elite SFF computer, i would get an error saying: An error occured during the installation of the device. The name is already in use as either a service name or a service display name.
Appearantly some other device driver already created a service called HECI , with a service Display Name of “Intel(r) Management Engine Interface”
To resolve this, i had to delete this service in the registry, reboot, and after that the driver would succesfully install.
Specifically i had to remove the registry key HKLM\System\CurrentControlSet\Services\HECI
This key actually points to the correct driverfile (heci.sys) and also the correct version of the file.
But when i re-install the driver for the IME device, it creates a different service:
HKLM\System\CurrentControlSet\Services\MEI
With also a DisplayName of “Intel(r) Management Engine Interface”
So the other service called HECI is preventing the MEI service from being created, and thus, the driver from being installed.
How did this HECI service get in there in the first place?
Maybe it was already in the capture image i made with sccm?
I might have to remove it manually from this capture image then.
Or, maybe i can edit the .inf file of the heci.sys driver and change the DisplayName of the Service so they won’t conflict?
To be continued…

SCCM, error importing Windows 7 only- drivers to repository

August 2, 2011

On a Windows 2008 (Non R2) based SCCM 2007 SP2 server, importing a Windows 7 -only driver fails with error: ” The selected driver is not applicable to any supported platforms”.
This happens because the OS type for Windows 7 is not properly recognized when reading the drivers .inf file.
See KB978754 for this, and the patch.