Hey SCCM Users,

i'm running SCCM 2012 R2 SP1 and having big trouble getting my OSD running again. I started with non-SP CU4 some months ago and everything was working fine. Then suddenly OSD stops working randomly always with following log entry (smsts.log client-side):

CLibSMSMessageWinHttpTransport::Send: URL: sccm.server.local:80  CCM_POST /ccm_system/requestError. 
Status code 503 returned

I googled my way around and often found answers about Date/Time Bug. I tried many things and sometimes it really seams to be Date/Time related. Well, now with SP1 i guess the logging was expanded and i get the following error:

CLibSMSMessageWinHttpTransport::Send: URL: sccm.server.local:80  CCM_POST /ccm_system/requestError. 
Status code 503 returned
pNext != NULL, HRESULT=80004005 (e:\nts_sccm_release\sms\framework\osdmessaging\libsmsmessaging.cpp,2054)
reply has no message header marker
...
...
failed to request for client
SyncTimeWithMP() failed. 80004005.
Failed to get time information from MP: http://sccm.server..local.

Well, it's obviously a time/date problem. But i don't get how to handle it. My client's BIOS time is correct. My SCCM Server's Windows-Time is correct. The smsts.log file (client-side) has a wrong timestamp. It's round about 10minutes in the future. When i boot up WinPE after the abort happened and copy the smsts.log to a smb share, the timestamp on the fileshare is many hours off real-time. 

Besides this i am often using HyperV-VMs to test deployment. In Hyper-V you dont have a BIOS to set up the time. The HyperV-Hosts time is syncing with the same ntp as the SCCM Server so they are the same. But i also get this error using Hyper-V VM. When i boot PXE many times then sometime it's working. Could be after 3 reboots or after 8 reboots...

Where can the client time be affected besides BIOS? Is there a setting somewhere regarding WinPE timezone or something?

Thanks in advance!

Edit: I just booted up the not-working client to OS and checked the log-file.

Created: 07/29/2015 00:20PM

Modified: 07/29/2015 00:27PM

Accessed: 07/28/2015 02:41PM

Okay...created and modified timestamp are tomorrow(?!). Inly the Accessed timestamp seems to be okay.

• Edited by Alexander Dmann 19 hours 5 minutes ago

Well the Status code 503 from the MP(IIS) mean that the service is unavailable.

Can you type this in a web browser: http://<ServerName>/sms_mp/.sms_aut?mplist (replace the servername wiht the mp name

You could also look at the IIS log to see why it`s returning 503

I tried your URL and it shows an xml with my MPs.

Can you try to deploy a new one now.

If it doesn't work you need to look at the MP server log and IIS log. 

Did you make any chance to the IIS on that server at some point?

I just checked the IIS Log and can't find any log entry with the client's IP address ( i looked for the ip-adress given in the smssts.log). Strange.

I booted the client to OS and tried your URL. Then IIS Log shows an entry.

So it looks like the client can not reach the Server during OSD?

But it's the same subnet, dns is working, i don't know what could be wrong.

Well you could start the TS and press F8 (need to make sure it`s enable in the boot wim) and you could test the connection from the winpe to the MP.

But clearly now that you found the issue just need to get the trouble shooting hat on and start having fun.

ping the sccm server's hostname is okay using F8 in WinPE.

There is no telnet available for testing HTTP...

Well, i've got hundres of "CCM_POST /ccm_system/request" entries with Status Code 200 in my IIS log from managed clients all over my net. But in WinPE it's returning 503 and nothing shows up in the IIS Log.

I'm out of ideas

Any way you could reboot the management point that is having issue in the log  please.

After start a new TS and if it fails can you upload the entire smsts.log please

• Edited by Frederick Dicaire 15 hours 49 minutes ago