locked
Client can't stream successfully - "Unable to read data from disk" RRS feed

  • Question

  • Hi there

    Since a few days (starting with our migration to appv 4.6 and win7) we some times encounter a situation, where a new win7 client is not able to load all subscribed packages. The client just keeps loading.. Server log states:

    [2010-09-01 15:23:59.178] %servername% 1628 1420 SW_CacheHandle::LoadBlock 297228900 "Default Provider" %username% Shockwave_Plugin_11.xc_Package 2 41502 "Unable to read data from disk."

    Maybe it is relevant to know that our appv servers load this packages from a central cifs share.
    After a server restart this problems disappear, but reappear a day later...


    Hints are very appreciated!

    Christian

    Thursday, September 2, 2010 6:55 AM

Answers

  • I think the SW_CacheHandle::LoadBlock indicates that the sft file is being located and/or is trying to load blocks from the sft into memory.  Even if the client has 100% load, the server wants to pre-load the first 512KB into it's memory just in case.

    Several server errors are associated with potential issues in this area.  For example 41520 (not 41502) appears, often with 41584, for issues in accessing the share, folder, and file.  41584 by itself happens if it can see the file but has a permission error opening it.  41508 happens if the server thinks the share doesn't exist.  Meanwhile 41523 happens if the file opened doesn't look like an sft.  Also 41486 is a special case if the package was sequenced with 64KB blocks and the server isn't configured for that.

    You might recheck the log for these isues, but since the problem is transitory, isolation (if possible) by resetting either the server or the file server may make sense.


    Tim Mangan MVP for App-V and Citrix CTP
    • Proposed as answer by znack Tuesday, November 22, 2011 8:08 AM
    • Marked as answer by Aaron.ParkerModerator Monday, December 19, 2011 10:27 PM
    Tuesday, October 5, 2010 3:34 AM
    Moderator

All replies

  • Are there any authentication issues? The Content share should allow rights for the computer accounts of the Management Server/s, is there an auth issue with those computer accounts accessing the share?
    Thursday, September 2, 2010 8:15 AM
    Moderator
  • It is designed like you say (an AD group having access to the content share with the appv servers as members of this group)..

    Do you have an idea on "how to check for auth. issues"? as it works "almost allways", and after a server reboot..

    Or do I have to check the people responsible for our AD?

    The other strange thing: The problem seams to be with "SW_CacheHandle::LoadBlock" .. does this mean it's a caching problem?

    Thursday, September 2, 2010 2:31 PM
  • Hello,

    Is there enough memory on the server?

    /Znack
    Thursday, September 2, 2010 9:26 PM
  • What kind of NAS device are you using? I am wondering if your kerb ticket is not renewing.
    Friday, September 10, 2010 10:51 PM
  • I think so..
    I will check the VM stats to make sure

     

    - thanks!

    Thursday, September 16, 2010 10:00 AM
  • I don't know exactly.. Another group is responsible for this device..
    As far as I know it is a Celera Box.. Don't know the exact specs..

    Thursday, September 16, 2010 10:03 AM
  • I think the SW_CacheHandle::LoadBlock indicates that the sft file is being located and/or is trying to load blocks from the sft into memory.  Even if the client has 100% load, the server wants to pre-load the first 512KB into it's memory just in case.

    Several server errors are associated with potential issues in this area.  For example 41520 (not 41502) appears, often with 41584, for issues in accessing the share, folder, and file.  41584 by itself happens if it can see the file but has a permission error opening it.  41508 happens if the server thinks the share doesn't exist.  Meanwhile 41523 happens if the file opened doesn't look like an sft.  Also 41486 is a special case if the package was sequenced with 64KB blocks and the server isn't configured for that.

    You might recheck the log for these isues, but since the problem is transitory, isolation (if possible) by resetting either the server or the file server may make sense.


    Tim Mangan MVP for App-V and Citrix CTP
    • Proposed as answer by znack Tuesday, November 22, 2011 8:08 AM
    • Marked as answer by Aaron.ParkerModerator Monday, December 19, 2011 10:27 PM
    Tuesday, October 5, 2010 3:34 AM
    Moderator