Post reply

$Mft corrupt

  • January 18, 2006 at 3:39 am

    #166570

    Hi,

    on my return from my Christmas holidays I noticed that my database and transaction log jobs were apparently not creating any files, although the job reported success. Running the job generated a popup on the server...

    The file or directory F:\$Mft is corrupt and unreadable. Please run the Chkdsk utility.

    I checked back in the system event log and found an entry on Dec 30th

    The file system structure on the disk is corrupt and unusable. Please run the chkdsk utility on the volume F:.

    A bit of research on the internet reveals that the Mft is the master file table, effectively the index of the files on the disk. F: is my data volume, a RAID 5 array, so this is quite a scary error to receive 

    My first thought has been to make a copy of the database to my backup server, and backup the database to a different logical disk than F:

    The database system seems to be running OK at the moment, I have not detected any problems.

    Please does anyone have any experience/ suggestions on how to deal with this problem. Should i take the server offline and run chkdsk ASAP. Any help much appreciated.

    David

    If it ain't broke, don't fix it...

  • January 18, 2006 at 4:17 am

    #615418

    take the database offline ASAP- take a copy of the files to a new array or remote location and check the disk out

    problems with the disk could lead to corruption in your MDF or LDF file which would be a nightmare (you may already have corruption but have not accessed the page or extent with the bad bytes in)

    problems like this can be sympomatic of Raid controller failuer, incorrect SCSI or RAID drivers, Poor disks, Poor power supply or just corruption due to power failure.

    MVDBA

  • January 19, 2006 at 1:34 am

    #615695

    Hi Michael,

    thank you for that timely advice. Having taken some backups of the databases to another server, I rebooted the server, finding that Windows 2000 had automatically scheduled a chkdsk with /f option. This reported an error in the space allocation table, which it corrected. All my missing backups and transaction logs reappeared in the problem directory, and my backups started working again (I have written some log-shipping procedures which use this directory).

    I have taken the precaution of reducing the time for which transaction logs are held in the directory to 1 week, as I wondered if the volume of files might be causing a problem (the logs ship every 15 minutes between 6am and 11pm).

    I also checked the status of the RAID array controller, as I had replaced the backup batteries on the before-write cache just before Christmas, but there were no reported errors. The server is on a UPS, so power failure shouldn't cause a problem. I will take some steps to ensure we get early warning of such problems in future.

    David

    If it ain't broke, don't fix it...

Viewing 3 posts - 1 through 2 (of 2 total)

You must be logged in to reply to this topic. Login to reply