none
Server 2012 RDS - User Profile Disks - Errors during Logoff RRS feed

  • Question

  • I have set up a test Server 2012 RDS collection (Single Server for now) and implemented User Profile disks.

    I have two problems.

    First: My generic test user can connect and does successfully use the user profile disk as expected. However, at logoff, the system event log contains these errors:


    The error (NTFS 137) is: The default transaction resource manager on volume C:\Users\ts3.test encountered a non-retryable error and could not start.  The data contains the error code.

    The warning (NTFS 50) that concerns me is:

    It appears that the user profile disk is being "disabled" or "disconnected" before the profile data is completely written at logoff. What can I do to troubleshoot this?

    Second:

    Update: A post from Mike Connor on the following page: -LINK- solved the problem described below. 

    My administrative user always logs on now with a temporary profile. At the beginning, the UPD was working and mounting. That stopped working. In attempting to troubleshoot, I logged the admin user off and deleted the UPD disk file from the share. I remember it working again after generating a new UPD disk file in the share. Soon, it quit working again. I deleted the UPD disk file again from the share and ever since, it has never regenerated a new UPD and always logs on with a temporary profile.









    Wednesday, December 19, 2012 4:01 PM

All replies

  • What is hosting the UPD share?  It is another Windows Server 2012 or something else?

    Don Geddes - SR Support Escalation Engineer - Remote Desktop Services - Printing and Imaging

    Wednesday, December 19, 2012 6:47 PM
  • Thanks for replying!

    The share is hosted on a domain joined (same domain as collection) 2008 R2 Enterprise SP1 - 64 bit.

    The RDS server has "full control" security rights on the shared folder. "Everyone" and the RDS server have "full control" on the share.


    Did I do that?

    Wednesday, December 19, 2012 7:13 PM
  • What is hosting the UPD share?  It is another Windows Server 2012 or something else?

    @dgeddes, does it matter? Is it possible to host the UPD share on a samba server?

    I am having this(http://social.technet.microsoft.com/Forums/en-US/winserverTS/thread/5a4e6fb8-ad10-4ac1-b3ac-55d414b3fb17) problem. Can you take a look. Thanks.

    Wednesday, December 19, 2012 9:56 PM
  • There have been reports of issues when the UPDs are stored on SMB shares that don't support SMB 3 or non-Microsoft SMB servers.  A quick way to check is to deploy your UPDs on a SMB 3 share and see if you still have the same problem.

    I think the issue that Chris is having though is more due to a network/disk driver problem as the event ID 50 seems to indicate. 


    Don Geddes - SR Support Escalation Engineer - Remote Desktop Services - Printing and Imaging

    Wednesday, December 19, 2012 10:26 PM
  • Thanks @dgeddes. I have no problem with deploying UPD on a windows share. Do you happen to know if UPD can be stored on a NFS share?

    Chris, sorry for jumping into your thread.

    Wednesday, December 19, 2012 11:35 PM
  • No, it needs to be SMB and preferably SMB 3. 

    Don Geddes - SR Support Escalation Engineer - Remote Desktop Services - Printing and Imaging

    Sunday, December 23, 2012 5:53 PM
  • Thanks for replying again, Don.

    So you are suggesting I test storing the UPDs on a Server 2012 share? Done.

    I verified the template UPD was generated and the User UPD was generated. It still appears to be disconnecting the UPD before the system is finished writing all the profile data. Anything else I should look at?

    No change:

    Error messages.

    - Chris


    Did I do that?


    Thursday, December 27, 2012 3:31 PM
  • I'm guessing you don't have anything else for me to try then?

    Thanks for trying.


    Did I do that?

    Thursday, January 3, 2013 3:58 PM
  • Additionally, I have completely disabled UPDs for the collection. Both my "basic" and "admin" users log on and off now with no errors in the event log.

    User profile disks seem like they would be far superior to folder redirection, as folder redirection misses some files within the profile, which are either discarded after logoff or have to be copied in at logon and out at logoff.

    We are excited about the potential time savings the UPDs would seem to provide, as well as the lower complexity involved in configuration.

    However, with 25-30 users per RDS server logging in and out sometimes several times a day, we would not like seeing that many errors in the event log. Also, we are not certain that corruption of user profile data is not occurring when these errors are evident each and every time a user logs off.

    If you have other troubleshooting steps I can take, I would be glad to try them. I dread the call to MS support and the time it will take to work with them.

    Thank you.


    Did I do that?

    Thursday, January 3, 2013 8:51 PM
  • In my opinion, you really need to look at this from a networking/file system standpoint and I'm not really an expert in that area.  I know you don't want to do it, but a network trace and working with Microsoft Support to review it is probably what I would do next.

    Don Geddes - SR Support Escalation Engineer - Remote Desktop Services - Printing and Imaging

    Friday, January 4, 2013 12:33 AM
  • I opened a Support Incident with MS support. Let's see where this goes.

    Thanks, Don.


    Did I do that?

    Wednesday, January 9, 2013 10:59 PM
  • Hi Chris,

    Could you please let me know if you have any feedback from MSFT

    I am having the Exact same errors after each user logoff/UPD volume dismount:

    The Error (NTFS 137)

    The warning (NTFS 50)

    The warning (NTFS 140)

    No Raid involved , or GPO's set

    My next step will be getting a Procmon trace to check what happens on the logoff time

    I will Also try to enable few of the RDS related GPOs mentioned on this link below :

    http://microsoftplatform.blogspot.ie/2012/09/overview-of-all-new-windows-2012-gpos.html

    I.E:

    *Turn Off UDP On Client

    *Suspend user Sign-in to complete app registration

    *Select network detection on the server

    Many Thanks


    Friday, January 11, 2013 12:26 PM
  • djamsite1 -

    I have been out of pocket and today is the first day I've had available to work on this ticket. I will post whatever I learn from MS back to this thread.

    If you come up with a solution, please post it here as well. 

    Thank you.


    Did I do that?

    Tuesday, January 22, 2013 4:58 PM
  • Hi

    I have the same errors in my setup. I am running the environment on a VMware ESXi platform 5.1

    This is my system eventlog

    Friday, January 25, 2013 11:14 AM
  •   The vibe I'm getting from MS is that this is a "common" issue. They are working with one or two clients and trying to come up with a solution or workaround. They told me they would keep me posted on results. The last time I spoke to them was a week ago, Wednesday (Jan 23, 2013).

    Did I do that?

    Wednesday, January 30, 2013 2:22 PM
  • Hi Chris,

    Thank you for the updates,

    I have been in Touch with MSFT and the issue seems to be a known as you said probably a BUG ...

    There is no solution so far it might take some time

    I will Share a lengthy info on what i have done so far with no luck

    I have tried everything possible to find the cause,

    These can be eliminated as causing this error, BUT YOU COULD CORRECT ME IF I AM WRONG !!!

    =================================

    *Configuration issue (tried on different labs same result)

    *Network or drivers ( Does not make a difference, took network traces and there is not packet loss or issue with SMB/RDP/TPKT protocols)

    *Disabled all the NIC features (ipv6-Chimney,TaskOffload,Jumboframes,VMQ ..)

    *Disabled SMB 3//2/1 through the registry and PowerShell (few new option through the SMB module)

    *Disabled few system services that are called during the RDS connection (spoolsvc , VSS ...)

    *Used the minimum options under RDP connection (disabling the local resources, Experience to the lowest...)

    *GPO (Tried all RDS ones including the UserProfiles policies)

    *RDSPNF.exe (new RDS service for 2k12 ) seems to take longer time to act before the logoff

    =====================

    The GOOD NEWS: there is no corruption of the profiles i have tried numerous times using scripts (logoff/logon)

    The only thing to avoid the errors is Detaching the UPD for the specific user from the RDS host before the logoff.

    However it might corrupt the profile :(

    The conclusion is that the user profile closes faster than the VHD of the UPD is detached or the other away around

    which might bring back some bad memories (UPHCLEAN) windows 2k3/XP issues

    UPHCLEAN

    ===========================

    Workaround is to create a PowerShell/Batch  to detach the VHD before the VHD detach,

    But i don't think that is a good solution it might corrupt the users profiles and make thing worse

    Chris please let me know if you get any updates i will also

    Regards

    Friday, February 1, 2013 5:45 PM
  • Hi Chris and djamsite1,

    I am having exactly the same errors as you mentioned before. Did anyone of you already get a useful feedback from MSFT ?

    Regards

    Monday, February 11, 2013 1:06 PM
  • FoG_Sigi -

      I have not yet received  any new information. MS collected more logs from us last week, but they have not reported any progress on the issue. I will post back to the thread once I have information that could help.

    - Chris


    I never learned from a man who agreed with me. - Robert A. Heinlein

    Wednesday, February 13, 2013 7:05 PM
  • Hi Chris, djamsite1 and FoG_Sigi,

    I have the same problems on my Server 2012 Session-host´s. I hope MS have a KB in the future.

    Regard´s

    Karsten.

    Thursday, February 14, 2013 6:05 PM
  •  

    Hi Guys,

    Just wanted to update you on the case,

    Still no feedback yet from MSFT it might take longer than expected, it’s definitely a BUG :(

    I believe that we need to wait till SP1 ...

    Regards 

    Djamsite1

    Wednesday, February 20, 2013 9:26 AM
  • Hi Chris and djamsite1,

    I only want to say thank you for keeping us up to date ;-)

    Wednesday, February 20, 2013 9:48 AM
  • Hello all -

     I have been informed today that Microsoft is refusing to troubleshoot our issue because it is virtualized under vmWare ESXi. They require we reproduce the problem on a physical machine.

    Anyway, We are evaluating whether or not this is important enough to us to go to these lengths to implement.

    Good luck.


    I never learned from a man who agreed with me. - Robert A. Heinlein


    Thursday, February 21, 2013 9:38 PM
  • Just thought I'd jump in and say that I'm having this exact same issue on 2x Server 2012 machines (belonging to separate customers), both acting as standalone RDS servers. The user profile disks is located in c:\userprofiles on the same server so there's no issue with network shares, drivers, SMB versions, it's as simple a setup as you could wish for.

    The only slightly non-standard feature of these servers is that they are virtualised (both are inside a Server 2012 host, one with no other guests). There is enough space, good high-speed SAS disks, battery-backed RAID card on latest firmware etc. All kinds of tweaks were made to optimised Hyper-V performance according to various guides available too, so I can't think what else there is to blame.

    It just so happens that I am having VSS issues on both these servers as well which are causing backup failures or crash-consistent backups at best, and the support guys at the backup software provider are pointing to these NTFS/WinLogon issues as the reason for the failures in both servers. (I have the same backup software running on several other hosts with no issues whatsoever).


    Saturday, February 23, 2013 2:19 PM
  • Guess what, we're monitoring this post because have EXACTLY same problem on a single physical machine, the nearest thing to virtualisation we've got going on is that the affected machine is an RDS server. 

    Eagerly waiting an answer MS....

    Monday, February 25, 2013 1:27 AM
  • Chris,

    Don’t let Microsoft cast blame on VMware so easily.  It’s disappointing that MS plays that wild card so easily.  Obviously this issue is not unique to ESXi.  Microsoft supports their server products on VMware through the Server Virtualization Validation Program (SVVP). SVVP is available to all customers through their normal support relationship with VMware or Microsoft. Any level of support from either vendor (including per call support) is acceptable to access this option.

    For more information see:  http://www.vmware.com/support/policies/ms_support_statement.html

    Jason

    Wednesday, February 27, 2013 2:46 PM
  • Hello all -

     I have been informed today that Microsoft is refusing to troubleshoot our issue because it is virtualized under vmWare ESXi. They require we reproduce the problem on a physical machine.

    Anyway, We are evaluating whether or not this is important enough to us to go to these lengths to implement.

    Good luck.


    I never learned from a man who agreed with me. - Robert A. Heinlein



    Chris, I hope you haven't given up on this yet.  You should not have been given that message.  Please send me an email with your SR # and I will look into this.

    Don Geddes - SR Support Escalation Engineer - Remote Desktop Services - Printing and Imaging

    Thursday, March 14, 2013 12:53 AM
  • dgeddes -

      Thanks for the offer. I'm not exactly familiar with Forum protocols and functionality. For the life of me, I cannot figure out HOW to e-mail you. Is there something I'm missing? Is there somewhere I can click to PM you or send an e-mail to you without already knowing your e-mail address?

      Also, from the other direction, can YOU e-mail me without that information?

    - Chris


    I never learned from a man who agreed with me. - Robert A. Heinlein

    Friday, March 15, 2013 2:51 PM
  • It is my forum name at microsoft .com.


    Don Geddes - SR Support Escalation Engineer - Remote Desktop Services - Printing and Imaging

    Friday, March 15, 2013 9:40 PM
  • I have the same error (ntfs 137) on my Windows  RDS servers. This two virtual machines are working on three nodes Hyper-V cluster. I tried to store user disks on Windows 2012 via network (on the same cluster), or directly on Cluster Shared Volume (SAN disk) connected to both RDS servers, or directly on system disk. Always the same error on logoff. It's definitly a BUG.

    Monday, March 25, 2013 8:39 PM
  • Final word. 

    No help from Microsoft. If anyone hears about any hotfixes or patches that will correct this, please post back on this thread for everyone.

    Looks like they dropped the ball on this one.


    I never learned from a man who agreed with me. - Robert A. Heinlein

    Thursday, April 4, 2013 1:11 PM
  • I'm looking into this for a different incident. If I develop a solution I'll be sure to share my findings here!
    Thursday, April 4, 2013 3:58 PM
  • So I am in the process of deploying an RDS farm. Should I enable the User Profile Disk or not? I did try it in the test setup before hand and it corrupted the Administrator profile so bad I had to completely trash it. I have now started from scratch. I have RDS installed on one server running Windows Server 2012 on VMware and waiting to add it to another physical server.

    Has anyone had luck with solving this issue?

    I am onsite ready to pull the trigger as I have 18 thin clients to deploy.


    Tuesday, April 9, 2013 7:14 PM
  • I too have had this problem on a pure 2012 platform. Everything was fine until a bad shutdown one day and ever since then it hasn't worked no matter what we tried. We did remove the bad profiles (ones with .bak) in Regedit under HKLM\software\Microsoft\WindowsNT\CurrentVersion\Profilelist\ , gut that only allowed us to connect to the corrupted profiles again. It still will not update the changes to the UPDs. This solution has great potential, but without a fix for this it's not viable for a production environment.

    How about it MS? Are you going to fix this or not?

    Sunday, April 14, 2013 10:29 PM
  • Any News? Same issue here at two of our clients. In this case UPD is not useable...thanks.

    Friday, April 19, 2013 2:17 PM
  • We are seeing the same issue on a Windows Server 2008 R2 SP1 with Hyper-V machine, Windows Server 2012 Datacenter edition with a 5 node RDS Session Collection Farm.



    Monday, April 29, 2013 2:40 PM
  • Hi Chris!

    First off, I'd like to thank you for your persistence with this issue.  We are about to deploy a two-VM RDS farm, both servers running 2012 Standard using UPD's, and are having the EXACT same problem!

    I was just posting here to re-enforce the fact that this is a HUGE issue that needs to be addressed by Microsoft ASAP.  Thanks for the time you put into this, and thanks to everyone else for responding.  Makes me feel better that we are not the only ones experiencing this issue.

    Friday, May 17, 2013 2:49 PM
  • There are 3 issues here that we are actively tracking:

    1. The event ID 137 and 140 from NTFS that seem to indicate that we've lost data from the UPD when there has been no data loss.  To date, we do not have any cases where real data has been lost, but if you are seeing data loss from the UPD then please open a case with Microsoft support so that we can investigate.  It is likely that these events can be safely ignored if you are not seeing data lost, but I do not have a confirmation on that yet.

    2. Users reporting corruption of UPDs after an unplanned outage of the file server where the UPDs are stored OR the Session Host server that had the UPD loaded (system hang, bugcheck, etc.)  In all cases, once the servers are stabilized, the problems with UPDs are resolved.

    3. Admins logging on directly to multiple session hosts that are in collections where UPDs are enabled. This causes the UPD to be loaded for the admin session and any subsuquent sessions for the same admin account will fail to load the UPD.  User will be notified that a temporary profile was loaded.  This occurs when logging on to the host via RDP or by using VMconnect (Hyper-V console).  This is by design. 

    In all cases, please do not accept your case being closed if you are hosting your Windows Server 2012 on VMware, or reopen your case that might have been closed previously.  If the issue does turn out to be a bug, ask to have the cost of the support incident refunded.


    Don Geddes - SR Support Escalation Engineer - Remote Desktop Services - Printing and Imaging

    Friday, May 17, 2013 11:25 PM
  • Hi

    Im in the same boat, i have experienced a few profile corruptions becasue of this bug, ive had one just this morning. Now the user gets temp profile. My recovery process is this. Reboot both RD Servers and File Server [UPD Store] after hours, then restore profile from a previous version.

    Environemt is this;
    2x IBM Servers and a IBM DS3524 SAN. VMWARE 5.1 on both hosts. And Server 2012 for the RD Servers.
    My UPD is on a Windows SBS 2011 server. Majority of the time users login and out without issue, but so far over a few months ive had profile corruptions.
    If they are at the satrt of the day i cannot resolve and have to configure the temp profile for them to use until i can fix it after hours.

    Ive done a bit of research online and there a lot of people with same issue. Microsoft cannot wash their hands of this. We need a fix.

    Hopefully we can keep this thread alive and get a solution posted too.

    Wednesday, May 22, 2013 12:01 AM
  • Hello,

    same problems here.

    RDS-Farm, 3 RD-Servers in it. very huge problems with the UPD and corruptions. (3 of 3 from dgeddes list)

    My temp workaround: set 2 of the 3 servers to decline new connections. all users are now on the same server but thats not exactly my understanding of a working RD-Farm with loadbalancing.

    That might work for the next 4-5 weeks but more users are about to migrate to the new farm if this happens it's going to be ugly.

    Help will be appreciated!

    Thursday, May 23, 2013 12:09 PM
  • Hello,

    Same issue here. TS Farm with 2 RDSH Servers User Profile Disk on shared folder.

    All servers are virtual on Hyper-v 3.0. Everything works fine except event log error. No corruptions is detected to user profile disk

    Any updates?

    Friday, May 31, 2013 9:07 AM
  • Hello,

    here the same problem.

    We have 4 RD-Servers and 2 Domain-Controllers running on 2 ESX 5.1 servers with 128GB RAM each.

    Our temp. solution is to deaktivate the UPD - but the UPD is a great feature.

    Microsoft, please come up with a solution.

    Wednesday, June 26, 2013 9:18 AM
  • I have the same problem at a customer site. However, our UPD location that gives an error is on a Server 2003 front end for a Dell NAS.

    When I point UPD to the session broker host, as a test, it works fine.

    Setup:

    2 RDS session hosts - physical machines

    1 RDS session broker - ESXi guest (Server 2012)

    Great feature - if it works.

    Is Server 2003 not a valid choice for the file share? I have both the session hosts and broker as full control of NTFS and Sharing and even made those three servers local administrators on the 2003 NAS host. On the Session Host (temporary test location of UPD) I have only listed the NTFS and Sharing for the servers - did not add to local admin group.
    • Edited by CNY-IT-Guy Monday, July 1, 2013 2:01 PM update
    Monday, July 1, 2013 1:51 PM
  • My issue was that the UPD share did not have the proper permissions. Someone had removed system from NTFS. All is well now.
    Tuesday, July 2, 2013 5:49 PM
  • Seeing the same issue.  I think it can be fixed by setting the virtual disks to quick removal and disabling write caching - but I can't figure out how to make that change and make it persist.  If we could set the virtual disk to discover with Quick removal enabled, that'd be worth testing.
    Thursday, July 25, 2013 11:00 PM
  • We had to use the same "temp" solution as Flow.Ryan. Spent days trying to get it to work and we have Hyper-V 2012 Data center edition with Hyper-V VMs. Would like to use UPD but it does not work. Had to constantly battle with temp user profiles issues.

    Phillip Riley

    Friday, July 26, 2013 8:58 PM
  • Seeing the same errors.

    Does anyone know how to map the Volume GUIDs in the event to an actual file or device? (either VHD file or physical volume). The errors I'm seeing always reference the same Volume GUID, so my suspicion is there's something wrong with a specific (single) vhdx.

    Thx in advance!

    M.

    Tuesday, August 6, 2013 12:38 AM
  • To convert the UPD SID (file name) to username and vice versa check this out: http://blogs.technet.com/b/heyscriptingguy/archive/2010/10/12/use-powershell-to-translate-a-user-s-sid-to-an-active-directory-account-name.aspx

    Friday, August 9, 2013 8:49 PM
  • Hi,

    To help with the SID's, I also wrote this little script to generate an overview of all the UPD files with the corresponding users. It's available on TechNet Gallery here:

    http://gallery.technet.microsoft.com/Retrieve-usernames-for-a-94780a9e#content


    Kind regards,
    Freek Berson
    The Microsoft Platform
    Twitter
    Linked-in
    Wortell company website

    Friday, August 9, 2013 9:05 PM
    Moderator
  • Does anyone know if this UPD problem still exists in 2012 R2 ?  Hoping perhaps someone has installed the Preview and tested this.

    Thanks - Tom


    tgoddard

    Friday, August 9, 2013 10:34 PM
  • Same problems here.  What was really bad was during testing, the Administrator accounts started to not be able to connect to their UPD's, so they kept using temp profiles.  I really hope 2012 R2 fixes these problems.
    Thursday, September 19, 2013 3:07 PM
  • It looks like Microsoft released a hotfix for this issue:

    http://support.microsoft.com/kb/2896328

    Thursday, November 14, 2013 6:57 AM
  • Just tried it.

    Unfortunately doesn't seem to work :-(


    Gil Gross | Project Manager & Technical Consultant | G-Net Network Solutions | www.g-net.co.il

    Thursday, November 14, 2013 12:21 PM
  • Hi Guys

    I have only one session host with local UPD and I'm getting the issue too. What occurs with user profiles if I turn off the UPD features ?

    Tuesday, November 19, 2013 2:43 PM
  • If you diable UPD then Profiles will be stored locally on server.

    You can find them in the users folder.

    Or you can set up roaming peofiles (not reccomeded) and or folder redirection.


    Gil Gross | Project Manager & Technical Consultant | G-Net Network Solutions | www.g-net.co.il

    • Proposed as answer by gjohnson63 Tuesday, April 22, 2014 3:58 PM
    Tuesday, November 19, 2013 2:49 PM
  • I understand that the profiles will be stored locally, but I just want to make sure that the existing profiles would be saved on their "actual state" . I fear the users loose their profile customization
    Tuesday, November 19, 2013 3:16 PM
  • Same problem here. 3 RDS servers with a connection broker. the UPD are stored on a Windows server 2012 and all the rights of the shared folder are correctly.
    Thursday, November 28, 2013 9:57 AM
  • Same problem here. Tried several different options, no luck. Only thing working is local profile.

    Hope Microsoft comes with a decent fix, as it now renders the complete idea of a broker with multiple session hosts completely useless.

    Friday, November 29, 2013 7:51 PM
  • Any one with a solution?

    In our case the problem is only related to thick clients ( Windows XP SP3) and not with our thin clients ( WIndows 7).

    We will try to investigate thiqs problem

    Monday, December 2, 2013 1:18 PM
  • I opened a case with Premium support regarding ths issue. I was told that it is a know case and I will be updated when a solution is found. 3 Months later, still the same.

    This morning I ran a test enviroment on Windows 2012R2 and guess what still the same issue! It seems that Microsoft don't have a clue how to fix this issue.

    Microsoft, do you want to wake up and fix the issue?


    Gil Gross | Project Manager & Technical Consultant | G-Net Network Solutions | www.g-net.co.il

    Monday, December 2, 2013 4:08 PM
  • Microsoft,

    A lot of users with the same problem, you need to solve this issue asap!!

    Tuesday, December 3, 2013 8:07 AM
  • I also checked out these hotfix info (without installing it):

    http://support.microsoft.com/kb/2896328/en-us

    It seems to be the same symptoms. Who else installed this fix and who is helped with it?
    In the description i don't see any remarks about 'UPD'

    Also it is not good news to read that in 2012 R2 this isn't solved.

    What is the alternative? Switch off UPD and set the user profiles local on the RDS host under C:\Users?

    How to set a quota on these user profiles?


    www.itprovider.be

    Thursday, December 5, 2013 5:27 PM
  • Anyone a working solutions?

    Wednesday, December 11, 2013 3:24 PM
  • We too are having this issue!  Anything from Microsoft yet on a fix????
    Friday, December 13, 2013 7:24 PM
  • I hope this issue has not been forgotten about.  I am beginning to run into this issue where UPD's are not loading and instead temporary profiles are loading for our users.  It happens sporadically.  Initially i thought it was due to a profile disconnecting rather than logging off and then trying to log onto a different RDS host server, but that no longer seems to be the case.  Below is our server set up:

    1 2012 management server

    1 2012 RDS broker server also handling licensing and RD web access

    2 2012 RDS Host servers

    1 2008 Server hosting the share UPD's are stored on

    This is a production environment with over 100 users accessing at any given time.  The issue is minor as of now, with only about 3 profiles having this issue per week, but they are starting to add up.  Has MS found a solution yet?

    Friday, January 10, 2014 2:20 PM
  • Yes, I hope this has not been forgotten about either!  This is causing us a lot of issues.  Microsoft, what's up???

    Tuesday, January 14, 2014 10:07 PM
  • All,

    After a call to Microsfot, we found a solution for our problem.

    The problem with the VHD only appears on Windows XP SP3 clients, not on the clients from windows 8 or windows 7.

    After 2 weeks support of Microsoft, the issue is solved.

    We have received a workaround from Microsoft.

    The VHD problem has been solved by excluding the VHD from the AV.

    Thursday, January 16, 2014 8:06 AM
  • Hi

    I also having issues with UPD on 2012 R2 which happen to be running in ESXi 5.1 U2.

    Got ticket open with msft and waiting for update to what the solution is.

    I believe there is hotfix for 2012 R1 servers however the problem appears to have been reintroduced in 2012 R2.

    Don't know if it works as all my servers are 2012 R2, here's the link anyway http://support.microsoft.com/kb/2896328

    Carl

    Tuesday, January 21, 2014 4:49 PM
  • Everyone complaining about temp profile issues need to first figure out why they are getting it in the first place. There can be a number of reasons for the same. Correct way is to look for the following event in event viewer on your Session Host Server: 

    Log Name:      Microsoft-Windows-TerminalServices-RemoteConnectionManager/Admin
    Source:        Microsoft-Windows-TerminalServices-RemoteConnectionManager
    Date:          xx/xx/xxxx xx:xx:xx
    Event ID:      20493
    Task Category: None
    Level:         Error
    Keywords:     
    User:          SYSTEM
    Computer:      rds.domain.local
    Description:
    Remote Desktop Services could not apply a user desktop for a user account with a SID of S-1-5-21-2871297466-3009485072-1309590870-500. A temporary profile was enforced for the user. Verify that the user profile disk settings are correct. The error code is 0x800700B7.0

    The error code can be converted using ERR.exe tool from Microsoft and then you can find the problem.

    For example the above error means (STIERR_HANDLEEXISTS) i.e. a handle is already open to that VHDX file and cannot be mounted. Further troubleshooting based on error code.

    Wednesday, January 22, 2014 3:20 PM
  • Manish, don't try to be smart about solving this. This problem is here since UPD was introduced in WS2012 and apparently persists still in R2. This has been confirmed by plenty of people with various configurations.

    For example in our environment, the RemoteConnectionManager log does not contain any errors.

    To those who opened a ticket with MS - did you get a hotfix/resolution? (aside from Samon900, but we're not using AV at all).

    Tuesday, February 4, 2014 2:41 PM
  • Found this thread after deploying a W2012 R2 RDS farm with UPD. All servers in the farm are getting all the errors described in this thread. Is there any hope for a fix?
    Friday, February 14, 2014 6:38 PM
  • Same issue here. Can't believe that there is no reaction from MS. This thread is about over a year!
    Tuesday, February 18, 2014 9:57 AM
  • Exact same error for me aswell in our environment at one customer.
    They are using Windows 2012 R1 in a farm using two RDS servers.

    As we see it it's because the UPD is not released from RDS1 when the user is told to go to RDS2 when hitting the broker.

    Wednesday, February 26, 2014 12:44 PM
  • Same issue ....
    Friday, March 28, 2014 2:14 PM
  • Any news? We have a customer with the same problem. Has anyone tried the hotfix?

    Regards, Christoph

    Wednesday, April 9, 2014 9:00 AM
  • I applied the hotfix (kb 2896328) but it did not resolve the issues.  We don't see as many of these issues but we are still getting them periodically. All systems are either Windows7 or Windows8. 

    Anyone have any ideas?

    Monday, May 19, 2014 1:16 PM
  • Same issue and a real pain. Opened a case with Microsoft, they seem to have no clue about the issue or the way to troubleshoot it. On first contact without understanding the issue completely i was asked to install the hotfix from kb 2896328. now we are at the phase of "further" data collection. the same issue exists in 2012 R2 as well. Microsoft please fix it or at the very least admit it it is an issue and you cannot fix it or don't have the time to! Even though the company i work for is a gold partner, after about two weeks of doing nothing, i was told to open a case under PREMIER contract!well, i am still struggling with Microsoft to progress it.

    Wednesday, May 28, 2014 6:34 PM
  • Same issue server 2012 R2... Pretty frustrating that is a known issue with know hotfixes since our MS licensing cost just as much the servers.   
    Thursday, June 5, 2014 1:53 AM
  • Hello,

    Same issue here, with 2012 and 2012 r2.

    It can't survive any unplanned problem.

    Quite often, one of the RDP server keep the vhdx open/locked on the filer, and then bad things happen.


    Mathieu Chateau http://www.lotp.fr

    Monday, June 16, 2014 7:13 PM
  • Hi all,

    I am facing exactly in the same Issues. We have 5 RDS Server with 2012 and use UPD. We have application which try to write information to the UPD, when user is logging off. Data get lost, UPD are in user by a different RDS Server, or completely un accessible. Next log on will end in temp profile, which will not work. So sometimes we have to kick off 300 Users to reboot the server to lock off the UPD.
    This is a disaster. We have to react very fast. We could not got on like this.

    I did the following workaround to stabilize al little bit the situation.

    In des Collection Settings, a set the disconnection time to 16 hours, because my users normaly do not sign off.

    I disabled log on with temp profiles. So better the user could not log on, than having a not working profile at all.

    We open a business critical ticket with MS, but only data collection happened. This is very poor.

    What did all the other do in that post? did you switch back to roaming profiles? Is this worth it?

    Any idea will help! Thanks.

      

    Monday, June 23, 2014 8:22 AM
  • Thorsten, we're enduring the problem. However annoying and total failure on Microsoft's part (I can't see how this could get through testing and is present in 2 OS generations ...), it's still little less annoying than having to deal with roaming profiles.

    I've previously asked if others got any resolution, but noone cared to respond (or forgot about this thread).

    Best thing we came up is to spread sessions over as many RDSH as possible and reboot them regularly to lessen the impact. And regularly cleanup messed up entries in the profileslist registy entry.

    Wednesday, July 2, 2014 2:07 PM
  • We experience less to zero problems after an outage with our User Profile Disks when we use the option: "Store all user settings and data on the user profile disk". Other then when we use the other checkbox (even though all checkboxes underneath are checked). 

    Our test setup was, an RDSH 2012 r2 server(s), with an 2012r2 server holding the share.

    We first tried the "store all user settings" for the User Profile Disk, we logged several accounts onto our RDSH Server we then disconnected our server which holds the share to simulate an outage. After about 5 minutes we connected the server again holding the share. When the share comes back online, we logged off the users and logged them back in. The result was a valid profile for all the accounts including their last changes from before the share got disconnected. The exact same test was done with the option: "Store only the following folders on the user profile disk" with all checkboxes underneath checked, which gave us corrupt profiles.. We also did these tests by giving the server holding the share a reboot instead of an disconnect from network. This gave us the same result. Each test was done at least 2 to 3 times..

    I hope others can confirm this as well in their situation, as it seems to be a bug within that specific checkbox.

    Saturday, July 5, 2014 8:56 PM
  • We experience less to zero problems after an outage with our User Profile Disks when we use the option: "Store all user settings and data on the user profile disk". Other then when we use the other checkbox (even though all checkboxes underneath are checked). 

    Our test setup was, an RDSH 2012 r2 server(s), with an 2012r2 server holding the share.

    We first tried the "store all user settings" for the User Profile Disk, we logged several accounts onto our RDSH Server we then disconnected our server which holds the share to simulate an outage. After about 5 minutes we connected the server again holding the share. When the share comes back online, we logged off the users and logged them back in. The result was a valid profile for all the accounts including their last changes from before the share got disconnected. The exact same test was done with the option: "Store only the following folders on the user profile disk" with all checkboxes underneath checked, which gave us corrupt profiles.. We also did these tests by giving the server holding the share a reboot instead of an disconnect from network. This gave us the same result. Each test was done at least 2 to 3 times..

    I hope others can confirm this as well in their situation, as it seems to be a bug within that specific checkbox.

    Looking at the above tests, the best way in my test cases were to exclude the default folders (f.e.: \Documents) if you would like to redirect them instead of using them inside your User Profile Disk. Redirects can be done through GPO Folder redirection options or GPO registry combined with Environment Variables.

    Even after several attemps to simulate an outage, i get left with consistent profiles. I've now also included an hard reset of my virtual machine (that gave the same results as a disconnect of the network card). This was all done through using the: "Store all user settings in the user profile disk". The other option left us with inconsistent profiles. I hope others can confirm this as well..

    Looks to me that the type of connection towards the User profile happens different. You can see this when looking at the C:\users\<username> and subfolders if you use user profile disks. The 2 options show different symbols as well.

    Sunday, July 6, 2014 10:18 AM
  • Thanks for the tests Dennis, but in our case, we never used any other option beside the "Store all user settings in the user profile disk" and we still get the problems described in this thread.

    How many users are we talking about? From our experience, the busier the server, the worse of an issue it is.

    Monday, July 14, 2014 7:04 PM
  • Thanks for the tests Dennis, but in our case, we never used any other option beside the "Store all user settings in the user profile disk" and we still get the problems described in this thread.

    How many users are we talking about? From our experience, the busier the server, the worse of an issue it is.

    Well in that case, i don't suggest doing otherwise then using that specific option (saved us a lot of effort into fixing temp profiles after we've switched.) We have about 100 users at this point, this number will increase during the next few months. Our setup does seem to be stable now. Using fully patches 2012r2 RDSH and a 2012r2 fileshare.
    Monday, July 21, 2014 8:07 AM
  • We have the same issue here

    ours is set to "store all users settings and data on the user profile disk" and we still get all the errors in the event log

    It seems odd that MS would not come in with a solution here, as 'User Profile Disks' were a great idea, yet it seems that this problem is still around even well into 2012 R2

    Our environment is Server 2012 hosted on Hyper-v 2012 - has anyone experienced this issue on hardware / not a VM?

    if anyone has a solution, love to hear it!

    Thursday, July 31, 2014 6:42 AM
  • Any update on this? Banging my head against the wall
    Tuesday, September 16, 2014 8:39 PM
  • Also STILL having this issue, will be one year next month.  Have not encountered any profile corruption(knock on wood). We have always had the "Store all user settings and data on the user profile disk" enabled.

    Our configuration is one RDS Server 2012 R1 VM hosting all roles on a Hyper-V Server 2012 R1 core machine.  Profile disks are stored locally on the C:\ of the RDS server.

    Another issue I believe may be related is we are running wbadmin backups of the RDS 2012 host VM from the Hyper-V core server.  The wbadmin backups occasionally fail(maybe once a week).  I think the issue may be related as the failures either a while when a user is working late(still logged in) or forgets to log out(profile left open), need to monitor a bit more to verify which.

    Tuesday, September 16, 2014 9:22 PM
  • Yep, same problem. RDS 2012 R2
    Tuesday, October 7, 2014 2:36 PM
  • UPDATE: Opened a support case and got nowhere. Said they hope to have it resolved in the next YEAR. Judging by the beginning date of this thread I would say that is their standard answer. Pathetically bad.
    Monday, November 17, 2014 8:39 PM
  • Two years later, any progress made?
    Wednesday, December 10, 2014 6:47 PM
  • Seeing this on a session host (server 2012) hosted on a Windows 2012 R2 pServer using direct attached storage.

    Help soon?
    Friday, January 9, 2015 6:03 PM
  • When this problem occurs, for while I solved doing this:

    - do logoff from RDS (temporary profile)

    - while this, go to your server with UPD folder shared; open Computer Management > System Tools > Shared Folders > Open Files, and find VHDX file from your user; it probably still opened twice (by same RDS), then Close the files (VHDX)

    - come back to your RDS and try logon again

    If you need to know what VHDX file is, put this script into notepad, and save with name showObjs.PS1 , then execute it like "showObjs.PS1 \\myserver\UPD" or change "$args[0]" by "\\myserver\UPD" on text. It will show you a list.

    $UPDShare = $args[0]
    $fc = new-object -com scripting.filesystemobject
    $folder = $fc.getfolder($UPDShare)
    "Username,SiD" > export.csv
    foreach ($i in $folder.files)
    {
      $sid = $i.Name
      $sid = $sid.Substring(5,$sid.Length-10)
      if ($sid -ne "template")
      {
        $securityidentifier = new-object security.principal.securityidentifier $sid
        $user = ( $securityidentifier.translate( [security.principal.ntaccount] ) )
        $user,$i.Name -join "," >> export.csv
      }
    }
    $a = Import-Csv export.csv
    $a
    pause

    I hope have helped till Microsoft fix this BUG!

    Friday, January 23, 2015 1:19 PM
  • Just found this exact issue on our system.

    UPD's are shared off a 2012R2 server.

    The VHDX for the user in question did not show up in the shared files list, other users did, and they were all working fine.

    The file existed, but the system process had it open.  I could not force it to close the handle

    Friday, February 27, 2015 8:56 PM
  • Just a quick note:
    Same problem seems to occur with non-UPD implementations on RDS 2012.  I have a customer with 2 RDS 2012 R2 host servers. The customer does noet use UPD, but roaming profiles + folder redirection.  Sometimes a user reports he/she is being logged on with a temporary profile.  When I check, I notice that the fileserver that hosts the redirected folders and roaming profiles, has files open in write mode for this user, even if the user is logged off from the RDS host.  Just closing theese open files in local-computer-share-management is all it takes so that the next logon of the user will be normal. (without temp profile)

    Not clear why this happens, but users report that when it happens, their logoff time takes extremely long, like 10 mins...

    I was reading this thread hoping that UPD might be a solution for my customer, but it seems that UPD suffer from the same problem: some kind of improper closure after RDS session end.

    Wednesday, March 4, 2015 3:04 PM
  • We have been struggling with roaming profiles as desribed by hwelvaar and thought that  UPD  was the solution. - As for the roaming profile we were able to have a work around the temporary profiles that where created by installing a Windows 2008 R2 server and use it to store the roaming profiles instead of the Windows 2012 Server. 2012 and 2012 R2 seems to hold the files locked and release them after 10 - 30 minutes while the WIndows 2008 R2 release them within 1 -2 seconds.

    We are testing UPD and are also using the same Windows 2008 R2  to save the profile disk. So far the UPD are release and there are no error. We only get temp profiles when we logging on the RDS with an Xp xlient (I know it is very bad).

    We will be testing this week more and we update the post.

    Monday, March 30, 2015 12:39 PM
  • Been following this thread for a while hoping for an update.  I guess UPD are not to be used in Production at this time.
    Tuesday, April 14, 2015 3:23 PM
  • +1 here. Server 2012 environment hosted on vmware 5.1

    ntfs issues in log but no corruption or tempfiles yet.

    Tuesday, June 2, 2015 3:42 PM
  • Running Windows 2012 R2 in vCloud and have had issues with every single user profile creating a temp. I wouldn't advice anyone try it in production until there is a fix that brings consistency.
    Sunday, June 7, 2015 9:54 PM
  • We are also using vCloud combined with User Profile Disks. In our case, the corrupt profiles have been caused by during the backup of the VMs. Probably the snapshot caused a brief disruption, which makes the profiles corrupt.

    After we asked the vCloud provider to temporarily disable the backup the corruption problems were solved.

    As a workaround, we have schedule a reboot before the start of the backup, so that no UPDs be in use as the backup is running.



    Erik Zwep

    Tuesday, June 30, 2015 7:29 PM
  • There was some time since anyone posted here. Anyone got any news? I have 4 servers running a fully patched (late Nov 2015) 2012 R2 RDSH VMs on VMware and use a UPD on a single 2012 R2 fileserver shared. Even though if I disable logons to 3 of the servers and let everyone login to the first server I get the problem for some users.

    And just to be sure I was starting out fresh on the 1st server (the Only one the could login to) I had:

    • Rebooted the server
    • Cleared out all profiles from System and C:\Users
    • Cleared out all old garbaged profiles in:
      HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\ProfileList
      HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\ProfileGuid
    • Made sure no .VHDX-files were in use
    • Rebooted again
    Friday, November 20, 2015 1:05 PM
  • We have been having very similar issues across all our collections. In most cases when users are logged off, disconnected in what ever means, not necessarily straight after logging in, the system has an open handle to the UPD .vhdx file e.g.

    \Device\Mup\....\...\..\UVHD-SID-ID.vhdx

    which is open by the System process.

    I've also noticed, in our case, file locks to the same users fontcache SID: e.g

    C:\Windows\ServiceProfiles\LocalService\AppData\Local\~FontCache-S-1-....dat

    Locked by the svchost.exe DLL

    There are also System, lsass.exe and svchost.exe handles open to the users security Token, which suggests a possible driver memory leak, but not totally certain.

    Sometimes the users profile is removed from the RDS session, but on some occasions the

    c:\Users\<username> folder is still present. c:\Users\<username>\UserData is the mount point to the UPD and can be checked using 

    fltmc volumes

    which displays the \Device\ mounts.

    So it seems that something is holding onto the UPD .vhdx on logout which then when the user logs into one of the other servers in the collection, they are logged in with a Temporary profile.

    I would be very grateful if MS can sort this out as it has been going on for some time now.

    Tuesday, January 5, 2016 11:15 AM
  • We are in the same boat here with one customer... already opened a support case with MS support.
    After several attempts still not near a possible solution... one was to disable Credential Manager with GPO to no avail...

    It seems to be performance related. When more users are logged on to the same RDS server the failure rate is higher. When less users are logged on to the same RDS server, failure rate is lower and occasionally none existent.

    Current user load on RDS servers (with 4 vCPU and 16 GB RAM) is not more then 6 active users, all server utilizations (disk, network, mem, cpu) is very low, users only have access to Office and IE/Chrome.

    We have several similar RDS environments with heavier load that do not experience this UPD dismount problem...

    Tuesday, January 5, 2016 2:56 PM
  • I also am seeing this with the following scenario:

    2012r2 File Server

    2012r2 Collection #1 (All Folders in UPD)

    2012r2 Collection #2 (Selected folders in UPD, Redirected Docs/App Data)

    Seeing issues with both collections building temporary profiles.  RDS servers have typical lockdown with no C:\ drive access from the users.

    Collection#1: Event ID 1515 & 1511 (Windows cannot find the local profile, used a backup, etc)

    Collection#2: Event ID 1508 & 1502 & 1515 & 1511 (Windows unable to load registry, Profile Corrupt, Insufficient Security Rights, Using backed up profile, etc).

    Collection #1 Observations:

    c:\users has a "TEMP" directory with security/owner of one of the errored user accounts.  I see "Shortcut" folders that look to be linked to the VHDX drives mounted in Computer Management.

    When looking at the drive mapping for these folders, I see they are mounted to the c:\users\%username% path.  This is not the case with Collection #2.

    Collection #2 Observations:

    c:\users has several "TEMP, TEMP.000" users, most are linked to users with regular folders, not looking to be linked, however the drive is mounted in Computer Management for the user.  Also seeing folders with .Backup.000 as the extension.

    I have identified a user that was no longer connected (forced signoff) and verified the VHDX was still mounted.  I forced the dismount of the disk and the folder on C:\users\%username% still existed, which means that the drives are not properly mounting to the c:\Users folder.  Upon further inspection, the VHDX is mounting to the c:\users\%username%\UserData path.

    This is interesting to see so much issue with the same problem for years.  I'm about to pull the plug on UPD but just checking if anyone else has any thoughts/ideas.  I have installed the KB3092688-v3-x64 patch and rebooted.  Same issues exist.


    Blaze T. Lewis


    Tuesday, January 5, 2016 4:12 PM
  • This thread shows the ignorance on MSFT's side ... last response from a MSFT employee in this thread is from may 2013, almost three years old. This issue is still present as can be seen from the user reports.

    I recall a thread or an article where Microsoft stated this issue MIGHT be fixed in 2016, but it didn't sound too confident.

    Anyway, all I can say is that it's ridiculous that the ONLY official way of storing user profiles when using RDS collection with 2 or more servers (besides outdated and problematic roaming profiles) is buggy and broken for this long with little prospect of being fixed.


    • Edited by MarkosP Wednesday, January 13, 2016 8:05 AM
    Wednesday, January 13, 2016 8:05 AM
  • Inherited RDS environment from customer who changed their IT support. I opened case to MS with the same issue. MS tech has remotely monitored user sessions and mounted vhdxs, when users logs off the vhdx is unmounted and and there are no open sessions to vhdx. Still the user gets temp profile from time to time and as RichardB78 mentioned this happens when more users are logging in. Disallowed new connections to 2nd server (have 2 RDSH) and no temp profiles. When allowed new connections temp profiles started to show up. Now I have added 3rd RDSH with new collection (same apps but in new folder in work resources). And no temp profiles in this new RDSH. Hoped that this new server solved the issue, but... (when the shit hits the fan) needed to publish a software (same as other servers has) and now this software do not work on the new server (made a support request for the software and now they investigates this one) After one and half month struggling MS has escalated my case (yesterday)...
    Wednesday, January 27, 2016 9:58 AM
  • Thanks for keeping this thread alive. Is there actually anyone running a RDS Farm on 2012 R2 with more than 1 RDSH and have UPD working properly without temp profiles?
    Wednesday, January 27, 2016 12:47 PM
  • Yes and no. We did, for a couple of hours (farm with 2x RDS 2012r2). But then all profiles started to get corrupt.

    But, we did not have any TEMP profiles issues. (2 rds servers with each 20-25 users)

    Then I stumbled onto this: https://support.microsoft.com/en-us/kb/2933664

    And more specifically, this: https://support.microsoft.com/en-us/kb/3092688

    We did not have network issues but we did have corrupted profiles. (but, again, no TEMP profile issues)

    Wednesday, January 27, 2016 9:49 PM
  • Did KB3092688 solve your issues or is it too early to tell?
    • Edited by MikaelJones Monday, February 1, 2016 1:59 PM
    Monday, February 1, 2016 1:58 PM
  • For me, too early to tell. End of next week we're going to do a full UPD test, it'll become clear very soon if it works or not. Fingers crossed.

    I'll get back to you on this.

    Monday, February 1, 2016 5:03 PM
  • For me, too early to tell. End of next week we're going to do a full UPD test, it'll become clear very soon if it works or not. Fingers crossed.

    I'll get back to you on this.

    I can tell you for sure that it won't permanently fix the issue..
    We have had an escaleted gold support ticket with Microsoft for months, and they keep asking us for dumber and dumber troubleshooting..
    We have created a powershell script which unmounts the VHDX every 5 minutes to "solve" the issue.. This Works pretty good, but my logging shows me that each day there are about 5-6 instances on the servers in question, where the script has to unmount the disk.. This does not happen on a single server setup! And we have a lot of farms.
    I'm afraid to say that UPD is a broken product, with no real priority from MS side.. It's too bad because the idea is there!
    Monday, February 1, 2016 5:32 PM
  • For me, too early to tell. End of next week we're going to do a full UPD test, it'll become clear very soon if it works or not. Fingers crossed.

    I'll get back to you on this.

    I can tell you for sure that it won't permanently fix the issue..
    We have had an escaleted gold support ticket with Microsoft for months, and they keep asking us for dumber and dumber troubleshooting..
    We have created a powershell script which unmounts the VHDX every 5 minutes to "solve" the issue.. This Works pretty good, but my logging shows me that each day there are about 5-6 instances on the servers in question, where the script has to unmount the disk.. This does not happen on a single server setup! And we have a lot of farms.
    I'm afraid to say that UPD is a broken product, with no real priority from MS side.. It's too bad because the idea is there!

    May I ask what the script exactly does? I have the exact issue at one of my customers as well and I was looking into creating a powershell script for this. I was wondering if and how you make the script check for open UPD disks that are no longer used in a session.

    Sunday, February 14, 2016 7:24 AM
  • Does anyone know if this issue exists when UPD are moved to a share on the same server? i.e. the problem is occuring due to some network congestion/issue that causes the sessions to disconnect and then not be able to reconnect to the same vhd

    We don't get profile corruption but when it happens it affects multiple users at the time of the ntfs 137 / 140 warnings/errors

    They tend to lose hours of work typing on a word report that doesn't save.

    If it does I think we will move to profiles stored on the RDP server which now has additional storage provisioned.

    Wednesday, February 17, 2016 5:00 PM
  • It's working good now, KB3092688 installed but also storing "the complete user profile" on the UPD.

    3 RDS servers in the farm and going good..

    Monday, February 22, 2016 9:12 PM
  • My issue is one of our 4 terminal servers 2012r2 is locking up the profile disks and not releasing them to be copied back to the store on our connection broker. Any ideas on this one
    Thursday, February 25, 2016 11:48 PM
  • harkx: Still working good with KB3092688 installed and "The complete user profile" set in UPD?

    TheShareonline: It would be great if you could post the script here! I guess you have some logic that checks that if the user is no longer online - disconnect the .vhdx?

    Wednesday, March 30, 2016 7:01 AM
  • MikaelJones: well, it was working fine until a couple of days ago. Event logs on the RDS servers started to get flooded with EventID 50, Source: NTFS, Delayed write failed. windows was unable to save all the data for the file <some file in profile>

    However, no disk/network issues. (hosted on SoftLayer hosting)

    After logging off and logging on again, it works for a while until those errors return. Next step is probably going to be: (1) abandon upd , or (2) retry with "Store only the following folders.." and only store the registry + roaming user profile.

    Tuesday, April 12, 2016 11:17 AM
  • An update, the problem remains. The complete user profile in UPD.

    Fileserver (on same VMware host as RD Session host) holding the UPD vhdx files, 1 RDCB, 3 RD Session hosts.

    When it starts failing, all users start to lose connectivity to the user profile. Iresulting in various application errors, being unable to save open word/excel files, etc..)

    So we logoff and logon the user but the UPD file stays opened on the  fileserver. (not viewable in Open Files list but only by using HANDLE.EXE from sysinternals) We have disabled the option to allow logging in with temp profiles so no more login for that user.

    So the system has the vdhx file opened (but user is logged of!) , the only thing we can do is reboot the fileserver, but after doing this the upd file gets "corrupted", the user can't login anymore. (we need to restore the UPD files)

    Last action taken: disable avast antivirus for the UPD folder holding the vdhx files.

    Wednesday, April 20, 2016 8:19 AM
  • Same problem and tried every trick in the book. updated vmware hypervisor to latests versions + updates.
    Changed UPD share to multiple (new) file servers. RDS 2012 R2 updated fully.

    Been working with MS support for over 3 months now and still no possible solution.... Our support case has been send to another ms support team for further analysis of trace files created (perfmon, handle, process monitor)... its been more than a month now without update or even a response...



    • Edited by RichardB78 Wednesday, April 20, 2016 9:24 AM
    Wednesday, April 20, 2016 9:23 AM
  • Hello,

    We experienced the same here on our RDS farm. Many files in use on our 2012R2 fileserver. We also use this fileserver as a printserver, and have a lot of trouble using the shared printers. Some shared printer connections are also  kept 'open' on the printserver.

    The result is that users cannot print to that printers. When they delete the printer and add it again, there are two printers visible.

    Today we introduced a 2008R2 server, to host the UPD's and the printers. And voila, problems gone.

    Also keep in mind that you cannot create a share for the UPD's on the RDS server itself, because then you get all kinds of trouble with your Internet Explorer like users that are unable to download/Upload, open PDF attachments etc.

    If you ask me there is something wrong with the SMB 3.02 protocol. Sometimes files are kept open in an idle state for ever.

    Hope I helped a little bit.

    Regards,

    Marino

    Thursday, May 19, 2016 9:48 PM
  • Really hoping someone actually has an answer to this, I also have a case open with MS on this for a large production environment where this is happening daily and we are not making much headway. We've installed every recommended update/rollup, removed AV, etc. 

    Ultimately we are pretty much leaning towards a daily server reboot to workaround the problem we really isn't ideal. Did anyone actually ever get this fixed? Thanks.

    Boseley

    Thursday, May 26, 2016 8:07 AM
  • Hello Boseley,

    Migrate the UPD's (and shared printers) to a 2008R2 server.

    Let's hope that this problem has been solved in the next version of SMB in server 2016.

    Regards Marino.

    Monday, May 30, 2016 10:52 AM
  • I'm also having the same problem with 2 2012r2 session host servers and 2012r2 file server - fully patched. It only occasionally logs a user in with a temporary profile but I can see that the profile disk is remaining locked even after log off.

    After following https://support.microsoft.com/en-gb/kb/3092688 without avail, The only resolution I have so far is to go into computer management and kill the affected SID. I also have a scheduled task to reboot running at 2AM on the server hosting the profile disks however this really isn't ideal.

    I wish Microsoft would sort this issue this is the first install I have completed with multiple session hosts on 2012r2 and decided to use profile disks with the thought process being it would be less configuration however its caused me more headache!

    <object id="kpm_plugin" type="application/x-KPMPlugin"></object>
    Wednesday, June 1, 2016 1:55 PM
  • Hey,

    I have some information to offer up. I'm not sure if it is related, but may help.

    We are using Server 2012 R2 for a DC and for two RDS servers, as well as User Profile Disks and redirected folders. It was ok for 5 or less users, but once either server was loaded with more than that, we were getting temp profiles, "Disk full or too many files open" error, slow app launch (up to 30 seconds) and long delays just opening mapped drives.

    I turned off UPDs, but the slowness and errors were still there.

    I did some tshooting and set up a mirror on the ethernet port that my DC was connected to, and used Wireshark to monitor traffic from the RDS servers. I was seeing a lot of "STATUS_INSUFFICIENT_RESOURCES" errors in the smb packets. I called Microsoft. For some reason they were disregarding the smb error as a valid piece of information.

    Did some research. It seems that SMB3 has a max connection limit from any single IP address, and also uses something call "long smb" queries which hold open connections longer than normal.

    I followed this article I found that talks about increasing the max connections:

    https://community.spiceworks.com/how_to/128605-rpc-issue-when-connecting-to-unc-root-share

    I think the “AsynchronousCredits“ value “1024” (default is 512) is the important key.

    I performed the reg changes 7 days ago (June 20th, 2016). The slowness is gone, and I've not had the "Disk full or too many files open error" since. I also turned UPD back on. I've not had a temp profile error either, however, I am still getting a lot of the following events:

    EventID: 51 disk "An error was detected on device \Device\Harddisk3\DR21 during a paging operation."

    EventID: 50 Ntfs (Ntfs) "{Delayed Write Failed} Windows was unable to save all the data for the file ."

    If this helps anyone, please let me know. I haven't had any problems yet, but I've not fully migrated my users from our 2008 RDS.

    Cheers!

    Tom

    Monday, June 27, 2016 4:31 PM
  • Ongwanada-IT, interesting! Now it's been Another week - still stable and no temp profiles? How many user are running in your RDS Environment?
    Tuesday, July 5, 2016 9:36 AM
  • Hey MikaelJones,

    As of today, things seem like they are working fine, no temp profiles. I only have 30 to 40 users cut over so far. When we are done, there will be around 100.

    One thing I did also do was disable logging in with a temp profile. I figured it was better to have a user call and say they couldn't log in the have them set their mail up again in a temp profile. That was a GPO thing "Computer/AdminTemplates/System/User Profiles - Do Not log users on with temprorary profiles".  I've not had one call regarding that either though.

    Still getting a lot of "EventID: 50 Ntfs (Ntfs)", "EventID: 51 disk" adn EventID: 140 Ntfs (Microsoft-Windows-Ntfs) warnings.

    The "EventID: 51 disk" I seem to get when a user logs on, but not every time, and we get up to 15 or so of them within the same second. The other ones I havent't really been able to correlate what/when they are being triggered.

    So far so good!

    Tom

    Monday, July 11, 2016 1:44 PM
  • Having the exact same issues.  Our setup:

    2x Connection Broker Servers in HA mode (configured for Collection Redirection), 4x Session Hosts, 1x File Server dedicated to UPDs, no Gateway Server.

    All Servers were updated with Windows Updates and RDS-specific hot-fixes before being put into production.

    When the issue happens, some users get a temporary profile. We have to find the Session Host where the UPD is still mounted and detach it in Disk Management, then delete the profile .bak registry entries from HKLM\Software\Microsoft\WindowsNT\CurrentVersion\ProfileList

    UPDs are configured to store all user settings on \\FileServer\Share. No excluded folders.

    Problem occurs randomly: different days and times, random users

    Monday, August 1, 2016 9:50 PM
  • Hi Tom - 

    I've been having issues with my session hosts not dismounting the UPD vhd upon logout and also some strange things with connections to the file shares. I was looking at the resolution regarding the SMB3 issue but using Procmon I cannot confirm the issue "Insufficient Resources" result. I'm tempted to try this solution anyways to see what results but I was wondering if you were experiencing specifically this issue in your environment along with the temp profiles // file share issues. Have had a case open with MS since early this year with no results so far. 

    Saturday, August 6, 2016 7:33 PM
  • Same problem for us too, really need Microsoft to fix this, it is hampering our attempts to roll out RDS.
    Monday, August 8, 2016 7:08 AM
  • We just began seeing this issue on our RDS farm after being in production for a couple of years without problems.

    Environment:

    2 2012 R2 RDSH in a Desktop collection

    2 2012 R2 RDSH in a Published App Collection

    1 RDS Broker server that also hosts the UPDs for each collection (2 different UPD locations- one for each collection)

    UPDs have worked flawlessly until recently (other than temp profiles used when you log in to both of the collection servers in the same collection at the console with the same admin account while "admining" the servers- obvious reasons i hope for everyone in this thread at this point). 

    I *believe* the problem started after the servers Ab-Ended when the hyper-v cluster they are on lost quorum and shut down.  Since then, we periodically see users login with temp profiles. We see their account logged in to ONE of the RDHS servers (in a given collection), but the UPD is still mounted on the other server in the collection.  If you still have a 2008 R2 server around with the old Remote Desktop Services Manager console on it (we do, just so we can send a message to EVERYONE on the RDSH server to notify of reboots, etc.) you can connect to your 2012/R2 RDHS servers and look at the Processes tab.  You'll see the csrss.exe, winlogon.exe, and the LogonUI.exe open on a session ID with no other processes. This, at least, is how we tell the UPD is still mounted and once we kill those processes, the user can login as expected as that unmounts the UPD as well.

    Since I'm pretty sure this started after an Ab-End, i'm wondering if a chkdsk might fix the issue...

    Oh, and I'll point out, our UPDs are on a deduped 2012 r2 volume, and even THAT didn't cause us issues... until now.

    Any current insight is appreciated.


    DLovitt


    Wednesday, August 10, 2016 9:36 PM
  • Hi All,

    We have the same issue on our RDS farm, but the problem has been traced to the Office version that is installed.
    The version we are using is the Office Click to Run from Office 365. It installed an service called Office Click to Run which keeps the UDP's mounted at logoff. Disabling the service helps with dismounting, but then Office won't work.

    So if you are still having troubles and you have Office Click to Run installed your only option is to wait. Microsoft is busy with the RDS and Office team on developing a fix. It is estimated for a october release.

    There is another forum on this issue which traced the issue to Office.

    https://social.technet.microsoft.com/Forums/windows/en-US/9de3a7bf-1cc7-47b6-b4d1-49d7955b6fc1/rds-2012-r2-office-365-shared-activation-upds-not-unloading?forum=winserverTS

    I hope it can point some of you in the right direction. Other RDS farms with Office from OVS does not have problems, so UDP should work fine now after installing all hotfixes and updates. Just not with Office click to run.

    Regards,

    Mats Nijman

    Thursday, August 18, 2016 9:09 AM
  • Hi All,

    First, good to know that we are not the only one experiencing this issue!

    We have several independent RDS farms affected by. Our main concern is the UPD not being demounted at user logoff resulting in temporary user profiles at next login.

    Mats Nijman, do you have any update on the Click-to-run solution?

    I can confirm that we have this issue on RDS farms with Office 365 licenced Office 2016 installation. But we also have RDS farms with normal Office 2016 licensed through SPLA (with no click-to-run service running).

    Thursday, September 8, 2016 12:38 PM
  • Hi All,

    First, good to know that we are not the only one experiencing this issue!

    We have several independent RDS farms affected by. Our main concern is the UPD not being demounted at user logoff resulting in temporary user profiles at next login.

    Mats Nijman, do you have any update on the Click-to-run solution?

    I can confirm that we have this issue on RDS farms with Office 365 licenced Office 2016 installation. But we also have RDS farms with normal Office 2016 licensed through SPLA (with no click-to-run service running).

    Hi taf_mynet - 

    Just want to clarify something in your post, your RDS farm with SPLA license installed does NOT have the issues? I am going to move one of my farms from 365 activation to SPLA to try to remedy this upd issue, was just wondering if you still see the issue with SPLA Office 2016 installed? 

    Tuesday, September 13, 2016 1:52 AM
  • Same problem here !

    Just a new setup , an lot's of probleme with UPD WHDX logoff still used...

    And for sometimes when the VHDX is weel closed, this is the SID in profilelist that is still alive and create temporary files !!!!!!!

    Friday, September 30, 2016 4:44 PM
  • Anyone tried the October 2016 Preview of Monthly Quality Rollup? https://support.microsoft.com/kb/3192404 Windows 8.1 and Windows Server 2012 R2 update history https://support.microsoft.com/help/24717/windows-8-1-windows-server-2012-r2-update-history Excerpt: "Addressed issue where the user profile disk (UPD) does not get unmounted when a user logs off. Therefore, users get temporary profiles and are not able to work with their own profiles during their next logon. The Event ID 20491 with a description of “Remote Desktop Services could not disconnect a user disk for the user account with a SID of <SID>. The error code is 0xAA.93” will be logged."
    • Edited by MikaelJones Thursday, October 20, 2016 10:11 PM
    Thursday, October 20, 2016 10:08 PM
  • Noticed that, and came here to ask the same question. You beat me by 22 hours and 17 minutes, Mikael.

    Alas, even if it solves it for 2012R2, no love yet for 2012 UPDs. And we had the problem first!

    Friday, October 21, 2016 8:30 PM
  • Noticed that, and came here to ask the same question. You beat me by 22 hours and 17 minutes, Mikael.

    Alas, even if it solves it for 2012R2, no love yet for 2012 UPDs. And we had the problem first!


    The patch roll up did not fix it in the environment I tested it on.
    Wednesday, November 2, 2016 6:40 PM
  • Hi all.

    I've 2 x 2012 R2 RDS servers with UPD enabled, storing UPDs on a 2012 share. KB3192404 didn't solve this issue for me either. 

    Except for this issue, UPD is great - but I'm under serious pressure to either solve this or go back to roaming profiles and then have to sort out Outlook OST cache file management (O365 users)...

    Attn Microsoft - please resolve this ASAP.



    • Edited by Jas at RMT Tuesday, November 29, 2016 5:52 AM
    Tuesday, November 29, 2016 5:48 AM
  • The October patch did not fix it in our environment either.
    Wednesday, March 22, 2017 7:56 PM
  • As other reported this is not enough. (my case owner at Microsoft also mentioned that rollup)

    We applied that rollup to both RDS servers and UPD-hosting fileserver.

    Something else I'm thinking about is creating a 2016 file server to house the UPD share.

    So far the problem remains but it does happen less often. (but still too much, when it occurs I need to restore the affected UPDs as they are corrupted)

    Monday, April 10, 2017 2:02 PM
  • Are there some Updates?

    I also already tried more than everything. I collected all workarounds and wrote a C# exe which automaticly detect temp-profils by using wmi, logoff users, detach vhdx files and than disconnects them from the connection broker (storing the vhd files).It removes all orphaned (vhd) devices from devmgmt.msc. It removes .bak entries (i also tried the gpo) and closes all handles of a the user. I tried to check vhds with checkdisk, I configured timeouts, I fixed dns entries by using /etc/hosts and many many more.... N O T H I N G   H E L P S ! ! ! N O T H I N G   H E L P S ! ! ! N O T H I N G   H E L P S ! ! !
    Wednesday, April 12, 2017 11:46 AM
  • Are there any updates?

    I also already tried more than everything. I collected all workarounds and wrote a C# exe which automaticly detect temp-profils by using wmi, logoff users, detach vhdx files and than disconnects them from the connection broker (storing the vhd files).It removes all orphaned (vhd) devices from devmgmt.msc. It removes .bak entries (i also tried the gpo) and closes all handles of a the user. I tried to check vhds with checkdisk, I configured timeouts, I fixed dns entries by using /etc/hosts and many many more.... N O T H I N G   H E L P S ! ! ! N O T H I N G   H E L P S ! ! ! N O T H I N G   H E L P S ! ! !

    Wednesday, April 12, 2017 11:47 AM
  • We're experiencing problems with UPD not correctly unmounted on Windows Server 2016 also.

    It's unbelievable to see that the problem hasn't benn fixed for years !

    Wednesday, January 24, 2018 2:19 PM
  • And now we add to UPD problems that MSTSC has been scaling broken on hi res displays since forever, and RDS multi-monitor support has been broken on W10 v1709 for 6 months.

    RDS is turning in to another of MS's red-headed stepchildren, despite them (finally) pushing forward with an HTML5 client. Even then, the HTML5 client is pretty lame for RemoteApps...doesn't simulate a local application UX at all. Very disappointing.

    Tuesday, April 3, 2018 7:03 PM
  • I inherited an issue from a client. I see the exact errors. They have around 160 users, using UPD and get the temp profiles logging in. Even a new user does. They have adequate resources for the session host and the file server. I seen same issue with 2016 when doing some testing. This client is on 2012 R2. As it seems and I just spent 30 minutes of reading, Microsoft has still yet to fix this issue.
    Tuesday, March 12, 2019 3:52 PM
  • Same problem with Windows 2016
    The problem is not resolved
    Unfortunately, I have no idea

    Thanks Micrsoft for the work ....

    greetings
    Michael
    Thursday, August 1, 2019 7:47 AM
  • Same fo us in RD 2019 !
    Friday, October 4, 2019 5:41 PM