File record segment xxx is unreadable

[bbacker]

I am currently experiencing strange behaviour on a Windows 2003 File
Cluster, with SAN Luns

Since a week we are experiencing more and more file issues.
- Not being able to delete files. Not structural all files, just
random files and sometimes directories.
- Not being able to copy files due to files not being accessible
(files are not locked)
- Free space not being returned to the total availlable free space
after deleting files.

Last days we decided to take the cluster offline to run checkdisks on
the local disks and SAN Disks connected to these nodes. Some SAN disks
are connected as mountpoints within the primary LUN to accomodate
expansion of data growth.

On most LUN's that are connected, when we run CHKDSK /F we get faced
with hundreds and hundres of the following error messages:

* File Record Segment xxx is unreadable

Some of the disks also return multiples of errors like:
* Insufficient disk space to hotfix unreadable file xxx

Can anyone tell me:
* what the possible cause of these errors are,
* how to solve it and how to prevent it from happening again,
* Is this a SAN, Harddisk, or OS level problem?

Anyone with any helpfull information is highly appreciated.

[John John - MVP]

See inline below:

bbacker wrote:
> I am currently experiencing strange behaviour on a Windows 2003 File
> Cluster, with SAN Luns
>
> Since a week we are experiencing more and more file issues.
> - Not being able to delete files. Not structural all files, just
> random files and sometimes directories.
> - Not being able to copy files due to files not being accessible
> (files are not locked)
> - Free space not being returned to the total availlable free space
> after deleting files.
>
> Last days we decided to take the cluster offline to run checkdisks on
> the local disks and SAN Disks connected to these nodes. Some SAN disks
> are connected as mountpoints within the primary LUN to accomodate
> expansion of data growth.
>
> On most LUN's that are connected, when we run CHKDSK /F we get faced
> with hundreds and hundres of the following error messages:
>
> * File Record Segment xxx is unreadable
>
> Some of the disks also return multiples of errors like:
> * Insufficient disk space to hotfix unreadable file xxx
>
> Can anyone tell me:
> * what the possible cause of these errors are,

Without a doubt the symptoms are all those of a corrupt MFT.


> * how to solve it and how to prevent it from happening again,

There is no way to solve this other than to reformat the drive and then
restore the files from a backup. I'm not sure how you would prevent it
from happening again because I am really unsure of the cause of the
corruption. Do you see any errors or clues in the Event Log?


> * Is this a SAN, Harddisk, or OS level problem?

If things were working fine and then these problems just started to
happen out of the blue then I would strongly suspect a hardware problem,
could be something as simple as bad cables or it could be that the
controller is defective. Here again I really don't know the exact cause
but my strong suspicion is that this is a hardware problem.


> Anyone with any helpfull information is highly appreciated.

Sorry that I can't offer more definitive help but the information might
steer you in the proper direction.

John