Going bald over this

[66945]

This is one of my worst experiences with SCCM.

 

I have a secondary site located in office X.

Then, I have a small remote office Y without any local site servers (DP, MP, etc) so all PCs there will pull from the secondary site server in office X.

 

Application & Software Update deployment and CPI works fine over S2S VPN from office X > Y.

 

However, package deployments just don't work.

I create various package deployments (some with, some without content) - they are all stuck at 'In Progress - Program Received'.

 

I see no errors in the logs at all - at best, the logs show me that 'nothing' is happening, i.e. the PC/agent sees the advertisement, and then does nothing (the next logical step is proceed to execute the package commands, which it doesn't) 

My boundary groups are set up correctly.

The very same packages, if I deploy to a different office (A with local site server on the same LAN), works flawlessly.

 

Can anyone help me before I lose all my hair...

[66945]

anyone able to help save whatever hair i have left?

[anyweb]

when you say 'they just don't work', what exactly happens ? if anything ?

[66945]

for example, with a 'required' package deployment

1. execmgr shows that agent received the 'advertisement'
2. in the correct flow of events, one would expect the advertisement to then kick off the installation/execution - but that doesn't happen
3. content is not downloaded
4. the deployment doesn't appear in software center either
5. reinstalling client seems to help - but only once; subsequently deployments require a re-installation each time

the agent definitely is talking to the MP and knows about the deployment; but it merely receives the deployment and doesn't proceed to do anything.

there are no errors in execmgr.

for a 'working pc', you might expect 10 lines of logs (for a successful and complete deployment)

but for agents in this remote office of mine, there are only 4 lines of logs (just an example) - and the remaining lines (steps) are missing with no errors at all (and the 4 lines of logs are 'correct' - i.e. they are what's supposed to appear for working deployments)

 

[66945]

Policy arrived for child program Install WinPcap	execmgr	3/26/2018 8:57:06 PM	4076 (0x0FEC)
Policy arrived for child program Install Wireshark	execmgr	3/26/2018 8:57:06 PM	4076 (0x0FEC)
Policy arrived for child program NI Count	execmgr	3/26/2018 8:57:06 PM	4076 (0x0FEC)
Policy arrived for child program Settings	execmgr	3/26/2018 8:57:06 PM	4076 (0x0FEC)
Policy arrived for child program Install Service	execmgr	3/26/2018 8:57:06 PM	4076 (0x0FEC)
Policy arrived for parent package PRI000F9 program *	execmgr	3/26/2018 8:57:06 PM	4076 (0x0FEC)
Raising client SDK event for class CCM_Program, instance CCM_Program.PackageID="PRI000F9",ProgramID="*", actionType 6l, value NULL, user NULL, session 4294967295l, level 0l, verbosity 30l	execmgr	3/26/2018 8:57:06 PM	4076 (0x0FEC)
Raising event:
[SMS_CodePage(437), SMS_LocaleID(1033)]
instance of SoftDistProgramOfferReceivedEvent
{
	AdvertisementId = "PRI20095";
	ClientID = "GUID:61A17DFE-EB8E-40C8-8C62-D7F5756AC1C1";
	DateTime = "20180327025706.984000+000";
	MachineName = "AZ-FS";
	ProcessID = 6076;
	SiteCode = "PRI";
	ThreadID = 4076;
};
	execmgr	3/26/2018 8:57:06 PM	4076 (0x0FEC)

As above, the policy arrive, and then nothing - client doesn't proceed any further, no errors whatsoever.

[GarthMJ]

Are your boundaries right? Can you see the content downloading? 

[66945]

I would like to assume my boundaries are correct -

Applications and Windows Updates are deploying (and downloading) correctly.

Only packages do the deployment get stuck and content does not get downloaded.

 

That said, it is occasionally possible to make a package download and install - i think it was by reinstalling client, but it doesn't work 100% of the time.

Which drives me bald.

[GarthMJ]

What type of boundaries are you using?

What boundaries match up with a problem client?

What is the IP and subnet mask of a problem client? 

[66945]

Before today (problem has been here for weeks) i was using IP address range and AD Site boundaries - for all offices - including those affected and non-affected ones.

As part of troubleshooting this issue, I decided to remove IP address range from all boundary groups and use only AD site boundaries.

My structure is something like this -

• Office A + Primary Site (1) + Boundary Group (Site is on same subnet as boundary group) 10.1.0.0/16
  • Office B + Secondary Site (2) + Boundary Group (Site is on same subnet as boundary group) 10.10.0.0/16
    • Problem Office C + Secondary Site (2) + Boundary Group (Office is of different subnet than the Secondary Site servicing this office) 10.20.0.0/16

[GarthMJ]

I would use IP ranges instead of anything else. This guarantees that everything is right, whereas IP subnet, AD boundaries, etc. don't.