InigoMontoya

that makes sense. So I shouldn't allow automatic client updates, I should manually push out new client updates myself. Good to know. Thanks Peter!

Hey all, I have a slight problem that I need some info on. I recently upgraded to SCCM 2012 R2 Version 5.0.7958.1501, which is working great. Problem is, the Automatic Client Upgrade Version that is being detected is still the old previous version 5.0.7958.1000. The client living on this primary site server/mp/dp/sup is still the old version 5.0.7958.1000. I assumed that would have been upgraded when installed the latest CU for the server itself. Is the "Automatic Client Upgrade" version determined by the Primary site's client version? thanks, Nick

I am having issues with a Lenovo W520. I am using SCCM2012 R2 and WinPE 5.0 and the latest USMT files from the Windows 8.1 ADK. performing a regular PXE boot with the same boot image works fine with WinPE 5.0 and the Lenovo W520. when I try to perform a user state migration on the same model laptop, it fails at "Initializing Hardware" because it doesn't load the NIC drivers. I am not sure how this is different from a PXE boot besides the fact that it is using the bios for drivers initially while downloading the boot image. I think it may have something to do with staging the boot image - but I am not sure how that could fail.. has anyone seen this before? yes I have tested that the user state migration task sequence works on other models... just something about this Lenovo W520. Please help!

So, while performing a user state migration, the computer loses power. The save state is on a network drive hosted by the SCCM server. what are my options for user data recovery?

I have around 50 configuration manager clients that all point to an old site - an old installation of SCCM 2012 SP1... I have a fresh installation of SCCM 2012 SP1 and I need to install new configuration manager clients for all machines in my environment. Can I do this simply by selecting the group of clients and going through the "install client" wizard making sure that it assigns a site and "always installs" the configuration manager client? what would be the easiest way of doing this? Also, would I have to do a similar process for EP clients? or will these automatically update policy once the new configuration manager clients are installed? many thanks in advance.

Here is an update - after a whole day of troubleshooting I have concluded that the only reason my task sequences were failing via Multicast is because they live on separate VLANs. So PXE booting using unicast works great - and I can use multicast as long as the machine is on the same VLAN as the PSP. This however, is not ideal. I need to be able to PXE boot across VLANs while using multicast. Anyone have any suggestions as to what needs to be done in my router to allow multicast across VLANs? Also - when I have a multicast session open even on the same VLAN it downloads extremely slow in comparison to the unicast session. Has anyone else noticed this behavior? Please advise

well, the issue is that it isn't even downloading the packages to begin with - it fails to open a multicast session because the MCSP isn't installing correctly on my DP. I will post the smstslog as well Attached is the latest SMSTS log - you will notice the error in downloading the file once it starts a multicast session, and then when it tries to check the hash it fails because it has no file to actually check against. Encountered error transfering file (0x800705B4). ApplyOperatingSystem 7/8/2013 12:10:09 PM 1436 (0x059C) Sending status message: SMS_OSDeployment_PackageDownloadMulticastStatusFail ApplyOperatingSystem 7/8/2013 12:10:09 PM 1436 (0x059C) Setting authenticator. ApplyOperatingSystem 7/8/2013 12:10:09 PM 1436 (0x059C) Set authenticator in transport ApplyOperatingSystem 7/8/2013 12:10:09 PM 1436 (0x059C) Sending StatusMessage ApplyOperatingSystem 7/8/2013 12:10:09 PM 1436 (0x059C) Setting message signatures. ApplyOperatingSystem 7/8/2013 12:10:09 PM 1436 (0x059C) Setting the authenticator. ApplyOperatingSystem 7/8/2013 12:10:09 PM 1436 (0x059C) CLibSMSMessageWinHttpTransport::Send: URL: VCG-SCCM1.vcg.local:80 CCM_POST /ccm_system/request ApplyOperatingSystem 7/8/2013 12:10:09 PM 1436 (0x059C) Request was succesful. ApplyOperatingSystem 7/8/2013 12:10:09 PM 1436 (0x059C) Download done setting progress bar to 100 ApplyOperatingSystem 7/8/2013 12:10:09 PM 1436 (0x059C) Succesfully performed multicast download from mcs://VCG-SCCM1.vcg.local/SMS_MCS/.sms_mcs?filePath=http://VCG-SCCM1.vcg.local/SMS_DP_SMSPKG$/VC100006&PackageFlags=134217760&packageID=VC100006&version=2 ApplyOperatingSystem 7/8/2013 12:10:09 PM 1436 (0x059C) VerifyContentHash: Hash algorithm is 32780 ApplyOperatingSystem 7/8/2013 12:10:09 PM 1436 (0x059C) Hash could not be matched for the downloded content. Original ContentHash = E568077A53B63FE644EDC46B1FB2E10017264E921E8D408347BE44FD449EA9E4, Downloaded ContentHash = smsts2.log

I was troubleshooting an issue with a hash mismatch during OSD and while going through the logs I realized that the hash's don't match because the package was never successfully downloaded via multicast! so there is no hash to check against! then I started looking into the mcsMSI and MCSSetup logs files - there are errors but I am not sure what is failing!? this, by the way is a fresh installation of the Multicast Service Point. first thing I tried to do is re install the role - but this has been unsuccessful. Please help me dissect these logs to figure out why the multicast service point is not installing properly! Thank you in advance! mcsMSI.log MCSSetup.log

I manage a unique group of users - 95% of them are developers who require local admin rights on their machines. So, I don't fully understand what you are saying. Lets say that I publish a package and it installs fine, but it installs another package first that runs a script to add registry keys - if I uninstall the application from a client through Add/Remove programs, then it would never undo what the original package installed. So if I create custom packages with scripts and or calling other packages first - then they will never leave the "Installed Software" section in software center. How would I configure detection rules for a package?

So, are you saying that if you install an application from Software Center, you also have to uninstall that application using Software Center? otherwise it will not detect that the application no longer exists on the machine? Normally (most end users) would use windows add / remove programs to uninstall software that is no longer needed or not working.

I realize that "Available Software" only shows new, not yet executed, installations - that is what I am saying is designed poorly. I have tested with a package and with an application, once I installed the application from Software Center, it will forever appear in "Installed Software" even if I uninstall the application from the client. Also, what do you mean detection rules? are you referring to the software inventory cycle schedule? If so, I manually ran all of those actions from the configuration manager client while the application was listed in "Installed Software" but in reality, it was uninstalled from the system. I guess that is the issue - if you uninstall an application that was installed through software center, will it ever show up in "Available Software" again? what Client data detection cycles have to run to detect that the application from Software Center is no longer on that system? and yes, F5 has become my best friend while testing in Software Center...

I am able to deploy packages fine, but when a client installs a package it is no longer displayed in the Available Software section of Software Center. So, if there is a problem with a program and it needs to be re installed, does that mean that users would have to click on "Installation Status" or "Installed Software" to find the package again and re install it? I have created a package that clients download locally and it runs a PowerShell script that starts the EXE for the program and then does a few other things. The software installs fine through the script, but in Software Center I can only see the package in the "Installation Status" section which says that it is installed. if I uninstall the software through add/remove programs, and run all of the actions within the configuration manager client - the package stays in the "installation status" section and still states that it is installed. If this is how Software Center is supposed to function, then Microsoft needs to change this. After you install a package or application, you should still be able to view it in the "Available Software" section unless you actually want to set the deployment to expire... Please help

Sorry, have been super busy here at work. I have read that if the PXE Service Point and clients reside on different subnets, that you have to define the DHCP options in order for clients to find the PSP. Is that correct? Yes, this was failing in a lab. I am almost positive that OSD failed due to corruption of packages and task sequences because of the machine constantly being "saved" and then "restored" constantly in Hyper-V. It was displaying behavior like it was looking for a TS that no longer existed. I know that removing roles and re adding them is a common troubleshooting step in SCCM, but no matter how many times I did that with the DP - it wouldn't get rid of that old TS. How would I manually go through the SCCM PKG folders and remove any traces of old packages / task sequences?

Right, I guess I am more concerned with what would happen if DA wasn't available for clients... Scenario: I have users who regularly connect up to Microsoft's Corporate Network which uses NAP. Those users for whatever reason, aren't up to date and Microsoft won't allow them into the network until they update the latest security patches and endpoint protection definitions. DA is not working on their PCs and when they go to update they cannot reach the SUP back at the office. now, I have endpoint protection set up to fall back to Windows Update after 0 hours if they cannot reach the SUP. BUT, for Windows updates - I would be getting a lot of calls about "not able to get updates on my computer" which, all they would have to do is select - get updates from Microsoft. But, that is a call that can be avoided. I want clients to be able to reach the SUP from the intranet and internet so if DA ever did go down, they wouldn't be without a MP,DP,SUP.. etc. I don't know where to start with this one. I have read about setting up a reverse web proxy with TMG for a SUP that will serve intranet and internet clients, but I don't know what is required for this, or if it is the best practice, and if it is - where do I start? I am looking for resources that I can study up on in order to allow clients to connect over the internet and the intranet.

In my organization, everyone is issued laptops to use as their computer. We have UAG DA, and we want to move it to DirectAccess 2012 - but that is later down the line. If DirectAccess is unavailable for Laptops, then I would like them to still be able to reach the DP, SUP, MP.. etc. Everyone takes their computers home and we all work from where ever we have an internet connection. What would be the best SCCM configuration solution for this environment?

http://systemscentre.blogspot.com/2012/05/system-cennter-2012-service-accounts.html half way down the page you will find the needed permissions for the network access account and others

To better state my last post: 1. Where are the logs that you requested located at? 2. Do I have to enable command line support to access these logs? 3. What do you think of my PXE enabling process that I outlined? I appreciate your help and I look forward to your response, Anyweb.

would I have to enable command line support during task sequence for this log to show up? I have never been able to find this log.. Also, does my PXE enabling process look good?

task sequence error 0x80070570. Task sequence failed to start. From the research I did it seems that either the task sequence is corrupted, files are missing, or it is pointing to a task sequence that no longer exists, or its a bad path all together to the task sequence. Like I said, No amount of uninstalling wds and removing the distribution point role and then adding again has been able to fix the problem except for uninstalling SCCM and then reinstalling. This is the process on how I PXE enable a distribution point: Enable PXE with SCCM 2012 1. Administration: Security: Distribution Point: right click on the SCCM server: Properties a. PXE: i. Enable all PXE check boxes (do not require password) ii. Only except PXE requests from the mac of the Internal VLAN Nic b. Multicast: i. Enable multicast with default settings 2. Configure DHCP to allow PXE a. 060 PXEClient (Not sure if needed) b. 066 Distribution point IP address c. 067 SMSBoot\x64\wdsnbp.com 3. Make a GPO for UDP ports 67,68,69,4011 Allowing for inbound and outbound 4. MAKE A NEW FOLDER CALLED SOURCES AND SHARE IT OUT! a. Under Sources, Make: i. Win8x64 ii. SUGs iii. Tools iv. Applications 5. Software Library: Operating Systems: Operating System Images: a. Add an Operating System Image – Install.WIM b. Right click the image – Distribute content to a Distribution Point c. Right click the image and under distribution settings, allow the package to be distributed via multicast. 6. Navigate to Sites: Select the VCG site: Configure Site Components: Distribution Point: Network Access Account: Specify the account that accesses network locations. 7. Software Library: Operating Systems: Task Sequences: Create a new task sequence a. Right click the TS and Deploy it to i. All Unknown Computers ii. VCG x64 Employee Computers

I can confirm that your method for prompting for a computer name works with unknown computers - my question now is: How do I get it to work with known computers that are already in a collection? Every time I try to PXE boot an already known computer, I get to the point of where it will prompt me for a computer name - but then the TS fails immediately afterwards. Also, this may be part of the issue. I am working in a virtual environment and after multiple times of saving and restoring the SCCM VMs, the distribution point seems to get corrupted. No amount of uninstalling wds and removing the distribution point role and then adding again has been able to fix the problem. So, know any reason why WDS/PXE/OSD would suddenly stop working without any changes? besides saving and restoring the VM?

I can confirm that the above method for prompting for a computer name works with unknown computers - my question now is: How do I get it to work with known computers that are already in a collection? Every time I try to PXE boot an already known computer, I get to the point of where it will prompt me for a computer name - but then it fails immediately afterwards OP please help!!