Strange issues with Application Deployment Evaluation Cycle

[toxaris]

Hello.

 

On some of our clients the Application Deployment Evaluation Cycle have stopped evaluate for some reason.

AppDiscovery.log, AppEnforce.log and AppIntentEval.log can be untouched for weeks and even months i.e the cycle havent been trigged since last timestamp on the loggfiles.

And if I manualy run Application Deployment Evaluation Cycle nothing happens.

 

As far as I can see, when a client have stopped running the Application Deployment Evaluation Cycle it will never start working again.

I have tryed verious things like reboot, restart ccmexec service, clear ccmcache, repair (install without removing it) the SCCM client.

Sofar the only thing that helps is running ccmsetup.exe /uninstall and after that reinstall it. After that is done the client works fine again.

 

I know that a client that have this issue does evaluate its machine-policy atleast, so the SCCM client isnt that dead.

 

Is there enyone that have seen this issue before?

 

//Daniel Persson, Sweden.

[Peter33]

Run this powershell script with admin rights on on of those clients and see if it returns something that is not 0.

If it returns one or more application names, delete and recreate the applications in the console.

I had a similar problem on 10% of our clients when the deployment cycle stopped working all of the sudden. This had been caused by a application which had been copied from another one in the console. I do not use the ability to copy applications anymore since then and the problem did not occur again.

[int]$countMissingDTs = 0

try
{
$Applications = gwmi -namespace root\ccm\clientsdk -query "select * from ccm_application"
}
catch
{
exit 1
}

$AppsTotal = $Applications.Length

foreach($App in $Applications)
{
$AppDT = [wmi] $App.__Path

if($AppDT.AppDTs.Name.Length -eq 0)
{
$countMissingDTs = $countMissingDTs + 1
$AppDT.FullName
}
}
$countMissingDTs

[toxaris]

Hello.

 

Its been a while since i wrote this post. Back then we found that this issue was releated to an application with alot of dependencys.

What we did then was to remove the application ofc, and also uninstall and install the CM agent on the clients and all was fine.

As a consequence, nowdays we rarely use the dependency method. Insted we use scripts to install multiple msi packages.

 

No here is the thing, we have encountered the issue again. This time in a quite bigger scale, atleast 1000+ clients.

We dont want to do the uninstall/reinstall CM agent again, thats just not a good sollution, aleast least not the manual way.

Note, the uninstall/reinstall CM agent method works this time also.

 

I have tryed your script this time, but it returns 0.

 

The only difference this time is that AppDiscovery.log have entrys as it should, other then that AppEnforce.log and AppIntentEval.log have no entrys for a many days and even weeks in some cases.

The issue came up again two days ago when we decided to distribute an application to 1300~ clients, around 1000~ havent installed the application yet.

 

 

Anyone have any idés?

[Peter33]

You might want to check the ciagent.log files on your Clients for Errors. This could be also caused by application Revision mismatches.

[toxaris]

For now I have created a schedule task that checks once a day if last write to AppIntentEval.log is more then 4 Days ago.

If, then CcmStore.sdf is deleted and the cm-agent healthcheck is initiated and that recreates the CcmStore.sdf and things seems to start working again after a while (10-20 minutes).

 

So far I dont know what casued this issue. But once again I think is releated to an application in someway.

In our enviroment we have 30k~ clients and most clients do install appliations as they should acording to other major distributions of applications like Adobe Flash.

 

I will post more about this when I have done futher investegation, but for now the storm is over atleast.