Hello all I have the problem in the title.  I have read a great many of these errors from these forums and have not found a fix to my issue.

I can resolve DNS just fine from winpe both short and fully qualified.  It works fine with network deploys but not with this media TS. 

I had this problem in the past but it was resolved with out explination as I was working on other issues with the deployment.  Mostly making mandatory deploys and changing settings in the TS.  But I cannot seem to get this one to resolve. 

for this Media TS I created a new collection and put both unknown computer types and the pc it self and assigned the TS as required to fix other issues I was having.  I want to check the server side logs to see if its rejecting it but I do not know which logs to check.

CLibSMSMessageWinHttpTransport::Send: URL: <server>.srf.local:80  GET /SMS_MP/.sms_aut?MPLOCATION&ir=&ip= TSMBootstrap 10/22/2013 8:36:57 AM 996 (0x03E4)
Error. Received 0x80072ee7 from WinHttpSendRequest. TSMBootstrap 10/22/2013 8:36:57 AM 996 (0x03E4)
unknown host (gethostbyname failed) TSMBootstrap 10/22/2013 8:36:57 AM 996 (0x03E4)
hr, HRESULT=80072ee7 (e:\qfe\nts\sms\framework\osdmessaging\libsmsmessaging.cpp,8936) TSMBootstrap 10/22/2013 8:36:57 AM 996 (0x03E4)
sending with winhttp failed; 80072ee7 TSMBootstrap 10/22/2013 8:36:57 AM 996 (0x03E4)
m_pHttpTransport->Send (0, 0, pServerReply, nReplySize), HRESULT=80072ee7 (e:\qfe\nts\sms\framework\osdmessaging\libsmsmessaging.cpp,5568) TSMBootstrap 10/22/2013 8:36:57 AM 996 (0x03E4)
CCM::SMSMessaging::CLibSMSMPLocation::RequestMPLocation failed; 0x80072ee7 TSMBootstrap 10/22/2013 8:36:57 AM 996 (0x03E4)
MPLocation.RequestMPLocation (szTrustedRootKey, sIPSubnets.c_str(), sIPAddresses.c_str(), sSiteCode, sAssignedSiteCode, sMP, sMPCertificates, sX86UnknownMachineGUID, sX64UnknownMachineGUID), HRESULT=80072ee7 (e:\qfe\nts\sms\framework\osdmessaging\libsmsmessaging.cpp,9583) TSMBootstrap 10/22/2013 8:36:57 AM 996 (0x03E4)
CCM::SMSMessaging::GetMPLocation failed; 0x80072ee7 TSMBootstrap 10/22/2013 8:36:57 AM 996 (0x03E4)
Failed to query http://<server>.srf.local for MP location TSMBootstrap 10/22/2013 8:36:57 AM 996 (0x03E4)

Any help?

• Edited by warzer Monday, October 21, 2013 11:42 PM updated details on error

I shall copy from above.

"I can resolve DNS just fine from winpe both short and fully qualified.  It works fine with network deploys but not with this media TS. "

it has an ip address or i would not be able to ping both short and fully qualified.  I have a feeling this is one of those many misleading errors the logs give where the cause is actually something else.

wanted to also add I setup a constant ping for an hr and did not loose a single ping to the MP

I also have this other issues http://social.technet.microsoft.com/Forums/en-US/7a688853-2ac0-4cd9-8f59-73a7213f6097/how-to-stage-task-sequence-prestaged-media-on-a-hard-drive-in-configuration-manager-2012-sp1?forum=configmanagerosd#52aadb6f-61fb-43e7-bc94-3e58fa26e851

to provide more info.  This new issue only started after i uploaded a new media TS wim file trying various settings like unattended checked or not and password or not.  each of the seem to give different errors.

• Edited by warzer Monday, October 21, 2013 9:04 AM

I have a feeling this is one of those many misleading errors the logs give where the cause is actually something else.

Errors are symptoms, not causes. Troubleshooting involves equating those symptoms to the root causes. The error isn't misleading, but your interpretation/analysis of it is.

How are you testing DNS resolution and what name exactly are you testing?

ping all components of system center.  ping all of our dc's, our name servers. and a few other random servers by name.  pinged the machine itself and pinged again forcing ipv4. all using names not ip.  are there any other ways i need to test dns?  I would like to test the MP service query from the client pc but i do not know how.  

from the display its right after the network setup.  Is there anyway to add a delay after the network setup and the next step? Maybe it just needs 1 or 2 more seconds to come up before querying the MP.  I checked our switch to make sure port fast is on, also watched the Nic light and it appears to be on during the time the connection fails.  But that is not a reliable test as all it would take is an imperceptible bink and that could be my issue.  Maybe add a ping command or something before the next TS?

Thanks for helping me with this.

Don

The only way to add any delay to the beginning of the process is with a pre-start command.

There is a new variable in R2 that may help (once you get to R2): SMSTSMPListRequestTimeout  (http://technet.microsoft.com/en-us/library/hh273375.aspx).

Setting SMSTSAssignmentsDownloadInterval and SMSTSAssignmentsDownloadRetry may help also (these are available in SP1+).

Note that you need to set these using a pre-start command since the problem is happening before the TS even starts running.

The only way to add any delay to the beginning of the process is with a pre-start command.

There is a new variable in R2 that may help (once you get to R2): SMSTSMPListRequestTimeout  (http://technet.microsoft.com/en-us/library/hh273375.aspx).

Setting SMSTSAssignmentsDownloadInterval and SMSTSAssignmentsDownloadRetry may help also (these are available in SP1+).

Note that you need to set these using a pre-start command since the problem is happening before the TS even starts running.

well, the variables were set and show up in the logs but it does not retry it MP connection and still fails at the same point.

What is different from a normal network pxe deploy vs the prestaged media deploy?  would another type of media deploy have the same failure as the prestaged? what other type of deploy uses the MP that I could try to provide more troubleshooting info?

Do you know if there is anyway to have it retry detecting/connecting to the MP during this process? 

I was looking at this:  but I am not running R2

Use this variable to specify how much time a task sequence waits before it retries to install an application after it fails to retrieve the management point list from location services. By default, the task sequence waits one minute before it retries the step. This variable is applicable only to the Install Application task sequence step.

I cannot seem to verify the default MP retry action in sccm 2012 sp1.  If it is supposed to retry it defiantly is not working as intended.

I was reminded of an older DNS issue we have that might be showing its head in this.  The DNS issue is that on older pcs like XP the first attempt to query DNS fails but the 2nd attempt succeeds.   I also have not been able to verify that we still have the issue anymore as every time I test it works fine.

Thanks

Don

To my knowledge, there's nothing more built in to help you in this scenario pre-R2.

I know some folks have added their own pre-start scripts that simply delay the start of the TS engine by adding a ping test and waiting for it to be successful.

I started working on this issue again and still have the issue.

here is what I have tried recently based on info in other threads.

Used a newer Nic Driver.  Added a 2nd MP.  Added prestart commands.

None of this allows me to get past the "Error. Received 0x80072ee7 from WinHttpSendRequest. TSMBootstrap 10/22/2013 8:36:57 AM 996 (0x03E4)"

With the 2nd MP in the mix it tries one fails then tries the second then fails with the same error.

Ping and lookups of both MPs works fine from the cmd prompt. 

Is there any way to query the MP in a prestart command to see if it functions there?

Maybe I am missing some requirement for advertising?  Does it need to be required?  Any special MP settings?  I am using these same MP for Pxe deployments for testing until we can get the pre-staged media to work.

Edit:  I had some other Thoughts.

Is there a way to statically assign an ip and dns info for the prestaged testing?

it is failing immediately after preparing network connections.  Looking at the logs I detects the nick correctly loads some settings then fails with unknown host.  As soon as I was able I pressed f8 during the process and started to ping the host and it fails until after preparing network connections.  Which is expected.  the problem is it fails so quick after I am not sure the dhcp server had time to respond before the failure.  As it is ping able as soon as that one step finishes.

I would also like to add a delay after network config before looking for the MP.  Any idea how this could be done?

• Edited by warzer 9 hours 49 minutes ago

Just for some sanity checks and follow-ups on statements you made:

- What prestart command did you add and why?

- When pinging from the command-prompt, you specifically mean that you boot a target system, wait for PE to start, press F8 to get a command prompt, and type ping <server>.srf.local (replacing <server> of course with the proper server name that you omitted above)?

- What does "I am using these same MP for Pxe deployments" mean? You don't choose an MP for PXE?

- Do you have portfast enabled on the switch port?

- When you do an ipconfig /all from a command-prompt on a machine booted into PE, does it show the proper info including your DNS servers?

- Deployments don't need to be required. That really has no bearing on MP/DNS issues.

- When exactly during the OSD process does it fail?

- Are you trying to use normal boot media or pre-staged media?

- What prestart command did you add and why?  Added SMSTSAssignmentsDownloadInterval =10 and ...downloadretry =10.  Why from the above in this post recommendations.

- When pinging from the command-prompt, you specifically mean that you boot a target system, wait for PE to start, press F8 to get a command prompt, and type ping <server>.srf.local (replacing <server> of course with the proper server name that you omitted above)?  CORRECT

- What does "I am using these same MP for Pxe deployments" mean? You don't choose an MP for PXE?

Instead of using prestaged, for testing since I cannot get prestaged to work, I setup pxe and deploy over the network that way using both MPs  Using the same task sequence this prestaged is based on and even using the same boot images.

- Do you have portfast enabled on the switch port? port fast is enabled.  As a test I had it turned off and the port too so long to come up, it was obvious to tell the difference.

- When you do an ipconfig /all from a command-prompt on a machine booted into PE, does it show the proper info including your DNS servers?  correct.  all information is listed and correct even DNS servers.

- Deployments don't need to be required. That really has no bearing on MP/DNS issues.

- When exactly during the OSD process does it fail? right after "preparing network connections" in the gui and in the logs it looks like right after the prestart vbs file exits with code 0 and it goes and tries to contact the first MP. The 2 lines above the errors listed are "Set authenticator in transport" "set media certificates in transport"

- Are you trying to use normal boot media or pre-staged media?  pre-staged as normal boot media works fine.

Forgot to mention this is also unattended.

I wanted to thank you again for helping me with this.

• Edited by warzer 6 hours 7 minutes ago added more info

I started working on this issue again and still have the issue.

here is what I have tried recently based on info in other threads.

Used a newer Nic Driver.  Added a 2nd MP.  Added prestart commands.

None of this allows me to get past the "Error. Received 0x80072ee7 from WinHttpSendRequest. TSMBootstrap 10/22/2013 8:36:57 AM 996 (0x03E4)"

With the 2nd MP in the mix it tries one fails then tries the second then fails with the same error.

Ping and lookups of both MPs works fine from the cmd prompt. 

Is there any way to query the MP in a prestart command to see if it functions there?

Maybe I am missing some requirement for advertising?  Does it need to be required?  Any special MP settings?  I am using these same MP for Pxe deployments for testing until we can get the pre-staged media to work.

Edit:  I had some other Thoughts.

Is there a way to statically assign an ip and dns info for the prestaged testing?

it is failing immediately after preparing network connections.  Looking at the logs I detects the nick correctly loads some settings then fails with unknown host.  As soon as I was able I pressed f8 during the process and started to ping the host and it fails until after preparing network connections.  Which is expected.  the problem is it fails so quick after I am not sure the dhcp server had time to respond before the failure.  As it is ping able as soon as that one step finishes.

I would also like to add a delay after network config before looking for the MP.  Any idea how this could be done?

• Edited by warzer Friday, December 06, 2013 2:33 AM

- What prestart command did you add and why?  Added SMSTSAssignmentsDownloadInterval =10 and ...downloadretry =10.  Why from the above in this post recommendations.

- When pinging from the command-prompt, you specifically mean that you boot a target system, wait for PE to start, press F8 to get a command prompt, and type ping <server>.srf.local (replacing <server> of course with the proper server name that you omitted above)?  CORRECT

- What does "I am using these same MP for Pxe deployments" mean? You don't choose an MP for PXE?

Instead of using prestaged, for testing since I cannot get prestaged to work, I setup pxe and deploy over the network that way using both MPs  Using the same task sequence this prestaged is based on and even using the same boot images.

- Do you have portfast enabled on the switch port? port fast is enabled.  As a test I had it turned off and the port too so long to come up, it was obvious to tell the difference.

- When you do an ipconfig /all from a command-prompt on a machine booted into PE, does it show the proper info including your DNS servers?  correct.  all information is listed and correct even DNS servers.

- Deployments don't need to be required. That really has no bearing on MP/DNS issues.

- When exactly during the OSD process does it fail? right after "preparing network connections" in the gui and in the logs it looks like right after the prestart vbs file exits with code 0 and it goes and tries to contact the first MP. The 2 lines above the errors listed are "Set authenticator in transport" "set media certificates in transport"

- Are you trying to use normal boot media or pre-staged media?  pre-staged as normal boot media works fine.

Forgot to mention this is also unattended.

I wanted to thank you again for helping me with this.

• Edited by warzer Friday, December 06, 2013 6:15 AM added more info

Status update:

I have not been able to get any further info as to why this is failing only in pre-staged unattended deployments.

It is very plausible that the NIC is not fully up when it requests the MP but I have been unable to verify.  After the GUI displays configuring network the MP is ping-able.  According to the logs right after this step the MP detect fails.  Any delays I have tried to add to the prestart have not succeeded.  I added a ping check that should last 30 sec but seems to fail to run. I still have not found a way to have it try more then once for a MP.  I think if it could recheck for the MP is would pass.

I am also unable to locate the requirements for prestaged unattended media deployments.

Still looking for assistance.

• Edited by warzer 8 hours 57 minutes ago

Status update:

I have not been able to get any further info as to why this is failing only in pre-staged unattended deployments.

It is very plausible that the NIC is not fully up when it requests the MP but I have been unable to verify.  After the GUI displays configuring network the MP is ping-able.  According to the logs right after this step the MP detect fails.  Any delays I have tried to add to the prestart have not succeeded.  I added a ping check that should last 30 sec but seems to fail to run. I still have not found a way to have it try more then once for a MP.  I think if it could recheck for the MP is would pass.

I am also unable to locate the requirements for prestaged unattended media deployments.

Still looking for assistance.

• Edited by warzer Tuesday, December 10, 2013 3:27 AM

So it defiantly is network delay at this point.  during the boot process the port is up and connected during the driver take over from the bios the port briefly goes down then comes back up and the system reboot.

I really need to know how to add a delay after configuring network step.  it looks like the delay needs to be about 1 second.  Any idea how I can add a delay?  I have tried adding ping to prestart but since it fails due to the network still coming up it does not add enough of a delay.  Does anyone have any suggestions for adding a delay?

• Marked as answer by warzer 6 hours 46 minutes ago

• Marked as answer by warzer Wednesday, December 11, 2013 5:38 AM

Hi Warzer,

In your prestart vbscript...does it simply contain the one line of "wscript.sleep 3000"?  Is there any kind of user indication that this is correctly waiting for the 3000ms?

I'm thinking about trying something similar since we've run into some sporadic issues with clients not being able to contact the MP during an OSD upgrade.  The computers seem to pull an IP address a lot slower in one of our building than the others and I'm hoping adding a command like this may help out.

Thanks again

it also sets 2 retry variables from SP1 that were added but they did not work on the initial MP check.  I left them in as having more retry's wouldn't hurt.  Those extra's are  "SMSTSAssignmentsDownloadInterval =10 and ...downloadretry =10"

You can see a dos prompt pop up when the prestart is run.  it also shows the prestart running in the logs.  But it does not show the wait.  I do not know for sure if its waiting the 3sec but it added enough of a delay for it to work.

Hope that helps