9:44 am in Uncategorized by nsienaert
Since a few months I’m working with SCCM 2012 and during my LAB experience I encountered some odd issues, which may sound stupid but quite blocking.
As I’m quite sure I will not be the only one on this planet that will bump into these I’ll post a small overview. [:)]
Gotcha 1: Application content could not be located
In the CIAgent.log I could see that my content could not be located.
How was that possible?
My Applications were stored correctly on the Distribution Point, I had no issues
with my environment (MP for instance) or what so ever.
Well the reason was Boundary Groups.
Boundary Groups are a new concept within SCCM 2012.
They are designed to simplify Boundary management.
To keep boundareis organized in logical containers, to avoid overlapping boundaries in migration scenarios.
BUT ALSO, now these groups are the primary object for content location so not longer the boundary itself.
Boundary Groups are added on the Distribution Point and on the State Migration Point.
Gotcha 2: User Device Affinity did not work
When starting to play with UDA I deployed a common scenario.
I had an application with 2 deployment types:
- One installing the MSI if the user is working on its primary device
- Another one installing the App-V version when NOT working on its primary device.
So what happend, I was 100% sure that my requirements were set correctly.
In AppIntentEval.log I could see that the MSI deployment type was always applicable even on a machine where a certain user was not the primary user…
The root cause was the application was deployed to a Machine collection.
If you target machines, primary device will always be true. BUMP.
Sounds logic of course but a mistake fast to make.
If you want to use UDA, you better target to User Collections. [:)]
Gotcha 3: Add tools to Windows PE
I like to add Trace32 into Windows PE for troubleshooting. If you open the properties of a boot image you will see that you are obliged to enable a prestart command hook, even if you don’t need that, you have to otherwise you cannot inlcude files…
So what I did was….. WScript.Quit(0)
Gotcha 4: Where do you import a machine?
With Beta 1, if you wanted to import machines into SCCM you had to navigate to the “User State Migration” node and select “Import Computer Information”.
Since Beta 2, you can only create a computer association on the “User State Migration” node.
So where is it now??
After digging for a while in the console I discovered it!
Right Mouse Click the device node and YES there you have it.
And yeah, if I watched the tabs more in detail I discovered it probably quite faster.
So lessons learned, use the new Tabs!
Gotcha 5: How do I troubleshoot the new App Model?
In this post I mentioned already some log files. With the new Application model there are quite some new log files. Here you have an overview of the most important ones and what they can tell you:
PolicyAgent.log: Check to ensure policy has been received by client
DCMAgent.log: Check log for Assignment ID (app + collection)
CIAgent.log: Evaluates CIs for App, DCM and SUM jobs
AppIntentEval.log: Contains highlevel information on applicability of each Application/Deployment Type.
AppProvider.log: Check Detection methods and install/uninstall specific failures
Also check these logs for infrastructure issues CIStore.log, CIStateStore.log, DCMReporting.log and CIDownloader.log.
All these test are done with SCCM Beta 2.
Till next time,