SOLVED User Collection application deployment failure - total components: 0

[Ratbasket]

Hi all,

TLDR: Server running 1906, all clients up to date. Applications deployed to device collections work fine. Content is distributed to all DPs.
When an application is deployed to a user collection it fails with error 0x0.
SCClient log gives error:
Failed to build instance path for ScopeId_96D2EB8A-0B4C-486E-99E5-555F1E2DD77A/Application_11c4d90a-0734-428e-ac9e-a036797b96a1 (Microsoft.SoftwareCenter.Client.Data.WmiConnectionManager at GetInstance) SCClient 04/09/2019 15:54:27 4 (0x0004)
and
Call to ExecuteMethod failed, WMI MethodName "ApplyPolicyEx", MethodClass "CCM_SoftwareCatalogUtilities", called by , error code -2016411097 (Microsoft.SoftwareCenter.Client.Data.WmiConnectionManager at ExecuteMethod) SCClient 04/09/2019 15:03:16 10 (0x000A)
and
Exception caught in ExecuteMethod, line 407, file X:\bt\1022889\repo\src\DataAbstractionLib\WmiDataProvider\WmiConnectionManager.cs - Type System.Runtime.InteropServices.COMException: (Microsoft.SoftwareCenter.Client.Data.WmiConnectionManager at ExecuteMethod)

**Edit**
Reading Microsoft's notes, the depreciation of the application catalog implies that the management point took over that role? We have 1 "primary" SCCM server with an additional server, "on site" and in the same boundary as all clients which only have a DP and site system role, does this site system server also need a management point?

Full background:

I'm a long time lurker of these forums, and have found really useful things here!

I've recently (well 6 months) changed role and now look after our company's SCCM. It is still very much a WIP as it was implemented a few months before I took over (and the guy who set it up has left : ( ).
It's only now that we have the need to deploy applications to collections which query AD groups, until now we've just deployed applications to device collections, which have been working fine.

When I deploy an application to a user collection, the users can see the application become available in software center, but when trying to install it errors immediately giving a 0x0 error code. Strangely, the total components = 0 and the download size is less than 1mb (when it should be about 50mb)

I thought this was initially to do with not having a application catalog role installed, but i know that 1906, which we've upgraded to recently, has depreciated the use of application catalog roles altogether. The only log which seems to show any errors seems to be the SCCLient log:



From reading around, I'm lead to believe it's unlikely to be an issue with boundaries? If it might be I can post a follow up/edit with how ours is setup. but don't want to overload this post with info.

Many thanks!

 

[Ratbasket]

I think the issue is with our environment, the way it is setup, the Primary site is sat in different boundary (lets call it B1) to most, if not all, user workstations/clients which are on another boundary (B2). When I've built a test workstation that sits in the primary site's boundary, the same deployment which wasn't working, now works (showing as having content, correct download size, and doesn't fail straight away)

I'll be reading up on this, but initial thoughts are:
a. Add primary site server to the other boundary B2
or
b. Add a management point to the site server in the B2 boundary

Or is there another option? Any suggestions welcomes thanks!

**EDIT**
Spoke too soon, looks like the workstation was actually in the same boundary as my other workstations, but strangely enough, it is having the same (original problem)
I noticed, when it DID work... the 'install' action triggered this in the log:

Getting a WMI instance for CCM_Application.Id="ScopeId_96D2EB8A-0B4C-486E-99E5....
but when it stopped working:
Getting a WMI instance for ScopeId_96D2EB8A-0B4C-486E-99E5-555F1E2DD77A/Appl....

Why is it doing this differently??
I've not done anything to the application deployment type or deployment, it's the same application.

 

[Ratbasket]

Ok. ignore all that. turns out when i was testing it on the new test machine, the client policy hadn't applied fully, which led me to think there's something up with that.
Revisited some MS documentation and turns out:
Enable user policy on clients
was set to no...

That was it. Enabled that option/set to yes, and there it was, the application now has content, and can be installed!

 

[rlgura]

Just ran into this exact issue and it was caused by the site server (also our MP) time being off by 13 minutes! Spent several days troubleshooting.