Sign in
Microsoft.com
 
Microsoft
 
 
 

locked
Help - New WHS PP3, 2 disks, seeing an NTFS error/File conflict RRS feed

 > 

  • Question

  • Question
    Sign in to vote
    0
    Sign in to vote
    After WHS running stable for the last few days (installed over xmas) I just today got home to see an 
    File Conflicts -> The file or directory is corrupt and unreadable
    error

    then when checking the server log I see some NTFS errors

     - this is a relatively clean install (No plugins, although SFU/interix is installed)
     - both disks were hw verified (read sector verification) before usage
     - this is an off-the-shelf system (HP EX490)
     - all updates are installed (Windows updates, inc. PP3, HP Update 1)

    so I worry that there is some underlying bug here?

    I'm somewhat unsure of next steps at this point.
     * just reboot 
     * force a chkdsk (C: D: ok or what?). NTFS errors are being reported for "D:" and "c:\fs\1ID"
     * sudden power off (vols marked dirty so then chkdsk will kick in!)
     * re-install (rather not)
    Tuesday, January 5, 2010 5:37 PM

All replies

  • Question
    Sign in to vote
    0
    Sign in to vote
    After WHS running stable for the last few days (installed over xmas) I just today got home to see an 
    File Conflicts -> The file or directory is corrupt and unreadable
    error

    then when checking the server log I see some NTFS errors

     - this is a relatively clean install (No plugins, although SFU/interix is installed)
     - both disks were hw verified (read sector verification) before usage
     - this is an off-the-shelf system (HP EX490)
     - all updates are installed (Windows updates, inc. PP3, HP Update 1)

    so I worry that there is some underlying bug here?

    I'm somewhat unsure of next steps at this point.
     * just reboot 
     * force a chkdsk (C: D: ok or what?). NTFS errors are being reported for "D:" and "c:\fs\1ID"
     * sudden power off (vols marked dirty so then chkdsk will kick in!)
     * re-install (rather not)
    I would start with running chkdsk /r on the hard drives in your server.  See the FAQ post:  How to check all the drives in your server for errors for details.
    Wednesday, January 6, 2010 2:54 AM
    Moderator
  • Question
    Sign in to vote
    0
    Sign in to vote
    See the FAQ kariya21 has linked as a start.

    But I'm concerned because you say you've installed SUA, which is not supported on Windows Home Server (nothing which requires desktop access is supported, technically). SUA makes use of advanced NTFS features for some of the functionality it provides, and some of those features (user-created reparse points, hard/symbolic links, compressed or encrypted files, etc.) are known to cause problems when used in the Windows Home Server shares. 

    If chkdsk doesn't reveal any issues, I think you will probably be best advised to remove SUA and/or use the server recovery DVD to perform a server recovery. 
    I'm not on the WHS team, I just post a lot. :)
    Wednesday, January 6, 2010 4:09 PM
    Moderator
  • Question
    Sign in to vote
    0
    Sign in to vote
    Understand the point.

    I'm hoping to minimize the risk by only using ssh/rsync & referring to all files using (effectively) UNC paths. In fact other than log files and the script or two it's read only.

    However the fact is that the storage solution is only viable if it's backed up, and my current provider requires ssh/rsync. If I can find another implementation, great. if not I'll have to revert back to a nas instead & sell up.

    unless there's another working option (cygwin is useless for this)

    chkdsk did find some ntfs errors. I might have thought DE problems could occur with any reparsing point issues, but not at the ntfs level. Fixed now will have to see if reoccurs. if so another drive scan will be needed. If clean, SW remains suspect.
    Wednesday, January 6, 2010 8:19 PM
  • Question
    Sign in to vote
    0
    Sign in to vote
    I should add that I ran chkdsk as per the Microsoft article. This found a couple of errors, after which the system was clean for a few days.

    However today I noticed "health" had gone yellow, and it reported that one of the drives was failing.

    The only "chkdsk" errors reported in the overnight run was a minor "free space marked as allocated". No "disk" errors in the event log, so this looks either like a left-over anomoly or some minor logical corruption rather than something hw related.

    Worrying, not in terms of this particular minor error, but in terms of whether there is doubt in the underlying system. running a "repair" now.
    Sunday, January 10, 2010 10:38 AM
  • Question
    Sign in to vote
    0
    Sign in to vote
    I should add that I ran chkdsk as per the Microsoft article. This found a couple of errors, after which the system was clean for a few days.

    However today I noticed "health" had gone yellow, and it reported that one of the drives was failing.

    Was it the same drive as before?  If so, you should start looking for a replacement drive ASAP.

    The only "chkdsk" errors reported in the overnight run was a minor "free space marked as allocated". No "disk" errors in the event log, so this looks either like a left-over anomoly or some minor logical corruption rather than something hw related.

    Worrying, not in terms of this particular minor error, but in terms of whether there is doubt in the underlying system. running a "repair" now.
    Sunday, January 10, 2010 6:08 PM
    Moderator
  • Question
    Sign in to vote
    0
    Sign in to vote
    Usually when a disk starts to fail there are driver-level/physical errors logged in addition to filesystem/application errors -- for example "disk" errors or "controller" errors. In addition whilst not tested for 2 weeks, prior to this both drives were thoroughly verified with a full media surface test prior to installation - this is certainly the case on every windows system I've had any disk problems on in the past (same with other OSs)

    I've also got no smart errors reported (although I wouldn't rely on that)

    Hence I rather suspect more of a sw error (leading to ntfs issues) rather than HW - a driver/filesystem/DE bug perhaps

    However
     a) I have an active online backup in any case :-)
     b) I will indeed do a media test again if any more serious issue occurs 

    Just FYI the actual error in this case was relatively mild --
    "CHKDSK discovered free space marked as allocated in the volume bitmap"

    As I said I have my eye out for some form of logical data corruption bug here.... something I never experienced with my previous NAS. Will pay close attention as this could go either way with either hw or sw at fault. Clearly as a fileserver data integrity is critical.
    Sunday, January 10, 2010 8:12 PM
  • Question
    Sign in to vote
    0
    Sign in to vote
    This problem has resurfaced - two very minor ntfs errors, but a more serious corruption issue today.
    I've added comments in the connect report at https://connect.microsoft.com/WindowsHomeServer/feedback/details/523262/whs-ex490-pp3-the-file-or-directory-is-corrupted-and-unreadable-new-install-no-plugins#tabs but I am *very concerned* at potential data integrity issues

    Basically it's the same issue as originally reported -- no sign of any disk/controller errors, but ntfs corruption cropping up out of the blue.
    Thursday, February 4, 2010 10:50 PM
  • Question
    Sign in to vote
    0
    Sign in to vote
    Try taking SUA off your server (via server reinstallation if need be).
    I'm not on the WHS team, I just post a lot. :)
    Thursday, February 4, 2010 10:53 PM
    Moderator