locked
App-V 4.5 SP1 client service crash RRS feed

  • Question

  • It's early days for diagnosing this one from my part, but thought I'd post what I'm seeing just in case anyone else has had the same errors, and can chip in...

    I'm seeing the odd client crash resulting in the service not responding and requiring a reboot to resolve; so far I've identified two separate packages which are/were running during or before the crash, but have not yet correlated any evidence to ascertain where/how/why it's falling over.

    Sample log file:

    [12/01/2011 15:28:24:835 SWAP INF] {hap=12C:app=AppName 6_0_8_24:tid=A30:usr=user.name}
    App shut down
    
    
    [12/01/2011 15:55:54:124 ???? ERR] {tid=DF8:usr=user.name}
    CVirtualRegistry HandleError 0x00000006 SeeIfSharedBlockIsLoadedAndAttach
    
    
    [12/01/2011 15:55:54:202 ???? ERR] {tid=DF8:usr=user.name}
    CVirtualRegistry::SeeIfSharedBlockIsLoaded IOCTL Failed for Guid SG-Sec9Load-{
    
    
    [12/01/2011 15:55:54:234 ???? WRN] {tid=DF8:usr=user.name}
    An error occurred while uploading the virtual environment (section: 9), rc 0F00D70C-0000003C
    
    
    [12/01/2011 15:55:54:296 ???? ERR] {tid=DF8:usr=user.name}
    CVirtualRegistry HandleError 0x00000006 UploadRegistryValue
    
    
    [12/01/2011 15:55:54:327 ???? ERR] {tid=DF8:usr=user.name}
    CSuite::InternalUpload failure  f00980c-      3c while uploading Virtual Registry
    
    
    [12/01/2011 15:55:54:343 SWAP INF] {hap=199:app=AppName 6_0_8_24:tid=DF8:usr=user.name}
    Elapsed time for upload: 0.219 seconds
    
    
    [12/01/2011 15:55:54:343 ???? ERR] {tid=214:usr=user.name}
    CVirtualRegistry HandleError 0x00000006 SeeIfSharedBlockIsLoadedAndAttach
    
    
    [12/01/2011 15:55:54:390 ???? WRN] {tid=DF8:usr=user.name}
    CSGApplicationManager::SysGuardCloseHandle DeviceIoControl failed = 0x00000006.
    
    
    [12/01/2011 15:55:54:405 ???? ERR] {tid=214:usr=user.name}
    CVirtualRegistry::SeeIfSharedBlockIsLoaded IOCTL Failed for Guid SG-Sec9Load-{
    
    
    [12/01/2011 15:55:54:405 TRAY ERR] {tid=AE4:usr=user.name}
    The Application Virtualization Client could not launch AppName 6_0_8_24.
    
    An unexpected error occurred. Report the following error code to your System Administrator.
    
    Error code: 4534C18-0F00980C-0000003C
    
    [12/01/2011 15:55:54:421 ???? WRN] {tid=214:usr=user.name}
    An error occurred while uploading the virtual environment (section: 9), rc 0F00D70C-0000003C
    
    [12/01/2011 15:56:04:681 SRVC CRT] {tid=FE0:usr=user.name}
    The Application Virtualization Client Service encountered an exception
    
    
    [12/01/2011 15:56:04:712 SRVC CRT] {tid=CC4}
    The Application Virtualization Client process is aborting (rc 04F01100-00000060).
    
    
    [12/01/2011 15:56:04:712 TRAY ERR] {tid=AE4:usr=user.name}
    The Application Virtualization Client could not launch AppName 6_0_8_24.
    
    The server threw an exception.
    
    
    Error code: 4534C18-0C701633-80010105
    
    [12/01/2011 15:56:09:803 SRVC WRN] {tid=CC4}
    process still connected to the listener

     

    -

    I've never seen half of these errors before, and the FTL tool isn't revealing much on these error codes, only the one on the 0C-3C code about paged pool, but client is XP, so not sure if it's related.

    App-V client: 4.5 SP1

    O/S: Windows XP Pro with SP3, IE7



    • Edited by Sizzl Friday, December 2, 2011 10:34 AM
    Friday, December 2, 2011 10:29 AM

Answers

All replies

  • Hello,

    See this article;

    http://blogs.technet.com/b/appv/archive/2007/09/17/softgrid-client-error-0c-0000003c.aspx (which you probably have read).

    It does give you specific suggestions that could be applicable.

    For more suggestions, see this post;

    http://www.virtualistic.nl/archives/550


    Nicke Källén | The Knack| Twitter: @Znackattack
    Friday, December 2, 2011 11:30 AM
  • Yep! The only relevant articles I've found were around Terminal Services / Citrix - multi-user environments and shared memory pools.

    The app isn't run from UNC or network drive (it's installed to MNT), so I think that discounts any other theory.

    It could indeed be memory related, so I'm probably better off trying to monitor for paged leaks from other drivers/processes than messing with the paged/pool values.

    Tuesday, December 6, 2011 4:18 PM
  • Without an exception code or a user dump of some sort it would be guess work. Obviously we have an exception based upon the logs. Have you tried a newer build?
    Steve Thomas, SSEE, Microsoft

    App-V/MED-V/SCVMM/Server App-V/SCCM/AppCompat
    http://blogs.technet.com/gladiatormsft/
    The App-V Team blog: http://blogs.technet.com/appv/
    The MED-V Team Blog: http://blogs.technet.com/medv
    The SCVMM Team blog: http://blogs.technet.com/scvmm/

    “This posting is provided "AS IS" with no warranties, and confers no rights. User assumes all risks.”
    Wednesday, December 7, 2011 12:19 AM
  • Everything listed in the above log was mirrored to eventlog, so there's not much more there... Would the process have likely created a crash (or user) dump on it's own? If so, where can I find that?

    Client is running 4.5 SP2; haven't yet tested SP2-HF2, or even considered steps required to go to 4.6 SP1-HF3.

    I'm hesitant to change anything on the App-V side until I've ruled out external influences causing paged memory saturation, given how infrequently it's happening (when compared against the remainder of the client estate).

    Wednesday, December 7, 2011 1:47 PM
  • OK, if the client is running SP2, go to this key in the registry:

    HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\SoftGrid\4.5\Shared

    Set the following values:

    DumpType - Set to 2 (DWORD)

    GenerateDumps - Set to 1 (DWORD)

    DumpPath - Set to a path you will remember

    - Then see if a dump is created when the issue reoccurs.


    Steve Thomas, SSEE, Microsoft

    App-V/MED-V/SCVMM/Server App-V/SCCM/AppCompat
    http://blogs.technet.com/gladiatormsft/
    The App-V Team blog: http://blogs.technet.com/appv/
    The MED-V Team Blog: http://blogs.technet.com/medv
    The SCVMM Team blog: http://blogs.technet.com/scvmm/

    “This posting is provided "AS IS" with no warranties, and confers no rights. User assumes all risks.”
    Wednesday, December 7, 2011 3:50 PM
  • Hello,

    SP2-HF2 (which you mentioned) could be applicable since it states that there is a problem relating to the client not releasing all handles, which could consume in memory.

    http://support.microsoft.com/kb/2507096


    Nicke Källén | The Knack| Twitter: @Znackattack
    Thursday, December 8, 2011 6:41 AM