locked
In the case that downloaded files have been deleted, what behavior do clients have in "BranchCache"? RRS feed

  • Question

  • Hi!
     
    I've two questions about "Branch Cache".
     
    1. What behavior does it have in the case like the below?
     
    0)Condition
      There are・・・
      ・Two clients(A,B) set-up in "Distributed Cache" mode
      ・One contents server(CS) set-up in "Distributed Cache" mode for SMB
     
    1) A client(A) has downloaded two files or more from the conetents server into the client's local folder.
    2) A client(A) has deleted one of the files that were downloaded in 1).
    3) A client(B) is going to download the file that has been deleted in 2).
     
    I think like the below.
    ・The client(A) has not the file but it has the cache information of the file.
    ・When the client(B) ask the neigborhood to know which client has the targeted file,
     the client(A) tells the client that the client(A) has the file.
     (but this is pseudo information.)
     
    But I could not imagine the behavior that follows the above.
     
    Can the client(B) get the its targted file or can not?
     
    2. Can we match the information of local cache to the state that the files are deleted?
     
    If in the case of 1 the client(B) can not get the file, How can we repair this disaccordance?
    The command "netsh branchcache flush" will make all of cache information in one client be cleared.
    Is there the method that can synchronize the state corresponding to the erased specific file?
     
    Any help would be much, much appreciated.
    Monday, February 22, 2010 1:28 PM

Answers

  • Hi

    1. If a BranchCache client likes to open a file, it has to access the original file on the original file server first. Then it gets back the hash of this file. With that hash, it will ask around (in distributed mode), whether somebody else in the branch office has a copy of this file. That means: If the file was deleted in step 2 of your description, it's not possible anymore for client B to get the hash of that file from the original server - ergo it would NOT ask around for that particular file in the local branch office.

    2. Don't break your mind about the caching - leave it as it is. Because nobody will ask about this file, it will retire at a certain time, so that the space can be used for more actual information.

    Does this solve your problem?

    Cheers,
    Andre
    • Proposed as answer by cal77 Tuesday, March 2, 2010 9:55 PM
    • Marked as answer by Miles Zhang Friday, March 5, 2010 6:45 AM
    Tuesday, March 2, 2010 9:54 PM

All replies

  • Hi

    1. If a BranchCache client likes to open a file, it has to access the original file on the original file server first. Then it gets back the hash of this file. With that hash, it will ask around (in distributed mode), whether somebody else in the branch office has a copy of this file. That means: If the file was deleted in step 2 of your description, it's not possible anymore for client B to get the hash of that file from the original server - ergo it would NOT ask around for that particular file in the local branch office.

    2. Don't break your mind about the caching - leave it as it is. Because nobody will ask about this file, it will retire at a certain time, so that the space can be used for more actual information.

    Does this solve your problem?

    Cheers,
    Andre
    • Proposed as answer by cal77 Tuesday, March 2, 2010 9:55 PM
    • Marked as answer by Miles Zhang Friday, March 5, 2010 6:45 AM
    Tuesday, March 2, 2010 9:54 PM
  • 2. Don't break your mind about the caching - leave it as it is. Because nobody will ask about this file, it will retire at a certain time, so that the space can be used for more actual information.
    Tuesday, August 3, 2010 8:17 AM