Volume snapshots disappearing?

Discussion in 'File Systems' started by Barkley Bees, Dec 17, 2008.

  1. Barkley Bees

    Barkley Bees Guest

    I have deployed two identical servers (Storage Server 2003 R2 SP2 x64 - Dell
    NF500 + MD3000) with DFS Namespaces and Replication and one shared folder on
    the D: volume. This folder (Users) replicates between the two servers.

    I have also configured Volume shadow copies on the D: volume of each server
    and assigned 3.8TB for the shadow data (total volume is 10.9TB). I have
    decided to keep the default schedule of 7am and 12pm for now.

    Anyhow, for the first week or so it was taking snaps with no problem. Then
    the other day I had about 10 pilot users copy their user folder data to the
    new server (total ~12GB). All seemed well with replication working correctly
    between the servers but then today I noticed that the volume snapshots are
    disappearing which is quite dismaying to say the least. For the last two
    days I am seeing the below in the event log:

    --------------------------------------------------------------------------------------------------------------------
    Event Type: Error
    Event Source: VolSnap
    Event Category: None
    Event ID: 25
    Description:

    The shadow copies of volume D: were deleted because the shadow copy storage
    could not grow in time. Consider reducing the IO load on the system or
    choose a shadow copy storage volume that is not being shadow copied.
    --------------------------------------------------------------------------------------------------------------------

    I understand that under high IO load (as the error message states) that this
    has been known to happen but this is with only 10 test users. We have
    another file server (2TB of data) with VSS enabled and accessed by ~500-750
    users that has never experienced this issue. It, like the new server, is
    saving the volume snapshots to the same volume as the data.

    I'm unsure as to what may be the actual cause here. Is it incorrect for me
    to have VSS enabled on both servers that are replicating to each other (I
    wouldn't think is related). Appreciate any advice or feedback from folks who
    may have experience this or similar.

    Note: I verified that disk write cache is not disabled.
    http://support.microsoft.com/kb/826936 (several other similar kb's on the
    matter but no clear resolution yet)
     
    Barkley Bees, Dec 17, 2008
    #1
    1. Advertisements

  2. Barkley Bees

    HAL07 Guest

    I have also this problem, and it seem to be a common problem on all our fileservers. (Windows 2003 and 2003 R2's. 32bit)
    This is not related to replication, but in general the filesystem that fails to grow.

    I've tried resolving it several times, but I've never found time contacting Microsoft on it. So if you find a solution, please
    post it! :)
     
    HAL07, Dec 19, 2008
    #2
    1. Advertisements

  3. Barkley Bees

    Barkley Bees Guest

    Sadly, it has still been occurring on a daily basis. Here's my plan off
    attack:

    1. As a first step today, I have installed the following patch to the
    secondary replication server:
    Installed - KB943669
    http://support.microsoft.com/default.aspx/kb/943669/en-us (will watch for
    results and update)

    2. If the above does not resolve the matter I will look at testing the
    following: http://support.microsoft.com/default.aspx/kb/826936/en-us.
    I will create a new registry entry "MinDiffAreaFileSize" in
    "HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\VolSnap" and give it a
    value of 3000 (3GB).

    3. Then finally, if I must, I will reformat the drive to use 16k cluster
    size (instead of the 4k default) as this is stated to be the best practice
    when it comes to using VSS - http://support.microsoft.com/kb/312067/.

    A very good reference here which I think pinpoints the problem:
    http://duitwithsbs.wordpress.com/20...o-avoid-loss-of-volume-shadow-copy-snapshots/

    The strange thing is that another separate server (mentioned in my original
    post) has been in use for over a year and is formatted using 4K cluster size
    and never had this VOLSNAP EventID 25 occur once (never lost any snapshots).
    Then again this new server is a fair bit larger in terms of capacity, using
    Dynamic disks to span several large volumes, and also DFS-R.

    Has anyone else been up this alley and found a valid solution? MVP folks
    also?
     
    Barkley Bees, Dec 22, 2008
    #3
  4. Barkley Bees

    Ronny Moreas Guest

    Ronny Moreas, Jan 14, 2009
    #4
  5. Barkley Bees

    HAL07 Guest

    Hi there,
    the following patch seem to fix some of these issues: http://support.microsoft.com/?kbid=940349
     
    HAL07, Jan 27, 2009
    #5
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.