Server 2k + Hard Disk Corruption.

[cybac]

Hi Guys,

I'm having a bit of a nightmare at the moment. I have a windows server
2k + SP4 environment in which I have 2 x 200gb Sata hard disks. The
first is split into 3 partitions, C: = 18.5gb and D: =83.8gb, E: =
83.8gb. The second is in 2 partitions, F: =93.1gb , G: =93.1gb.

Drive C = System
Drive D = Users (redirected My Documents and shared home folders)
Drive E = SQL + Apache Webserver
Drive F = Backups + Logs (AD logs)
Drive G = Recordings

Clients connect using XP with SP2 installed.

My problem is that every so often there seems to be some corruption
happening on one of the disks, the problem has occured on both disks,
on drives D, E,F & G. Small mercy that drive C has not been effected.

I get an error message that says the "file system on drive xx is
corrupt and unusable, please run chkdsk". I run chkdsk, it usually
deletes orphan files and rebuilds the file structure. After that
finishes the drive is useable boot sometime the data i need is either
gone or corrupt.

I have tried moving things and reformatting to get a new file system on
the partition, however over time the problem happens again. If it's not
dealt with straight away the server usually crashes and I have a bad
time trying to clean up the mess.

These drives contain a lot of user work and there have been a couple of
occasions where files have become corrupt and work lost. There are
quite a few shared folders.

Once the problem is fixed I check all the error logs and the system
seems stable again for a while.

I really would appreciate any suggestions as to what my next port of
call should be.

 

[Pegasus \(MVP\)]

Hi Guys,

I'm having a bit of a nightmare at the moment. I have a windows server
2k + SP4 environment in which I have 2 x 200gb Sata hard disks. The
first is split into 3 partitions, C: = 18.5gb and D: =83.8gb, E: =
83.8gb. The second is in 2 partitions, F: =93.1gb , G: =93.1gb.

Drive C = System
Drive D = Users (redirected My Documents and shared home folders)
Drive E = SQL + Apache Webserver
Drive F = Backups + Logs (AD logs)
Drive G = Recordings

Clients connect using XP with SP2 installed.

My problem is that every so often there seems to be some corruption
happening on one of the disks, the problem has occured on both disks,
on drives D, E,F & G. Small mercy that drive C has not been effected.

I get an error message that says the "file system on drive xx is
corrupt and unusable, please run chkdsk". I run chkdsk, it usually
deletes orphan files and rebuilds the file structure. After that
finishes the drive is useable boot sometime the data i need is either
gone or corrupt.

I have tried moving things and reformatting to get a new file system on
the partition, however over time the problem happens again. If it's not
dealt with straight away the server usually crashes and I have a bad
time trying to clean up the mess.

These drives contain a lot of user work and there have been a couple of
occasions where files have become corrupt and work lost. There are
quite a few shared folders.

Once the problem is fixed I check all the error logs and the system
seems stable again for a while.

I really would appreciate any suggestions as to what my next port of
call should be.

While I do not know what's causing your particular problem,
you might be able to extract some useful information from a
similar problem I recently had.

This was a brand-new HP server with two SATA disks. Shortly
after installing Trend Client Server Messaging, the machine crashed.
It subsequently booted up OK but Windows kept complaining about
a corrupted system disk. Running chkdsk either during a reboot or
under a Bart PE boot fixed numerous errors. It also caused the
loss of many files installed after the crash but before the chkdsk run.

After playing the chkdsk game for a week, I deleted the file system
on the system partition and restored it from an image I had previously
made. This was four weeks ago - the machine hasn't missed a beat
since then.

I assume you have checked the event logger for error messages.
I also recommend strongly that you install an automatic logger for
each and every boot event.

 

[Cybac]

Thanks Pegasus,

I'll install those boot loggers, and in answer to your questions yes I
have been checking the event log, unfortunately there is no error
messages until the actual corruption mesage telling me to run chkdsk.
If these boot loggers show up no further information I will re-install
the system area, unfortunately I do not have an image to install it
from so it would have to be a re-install of the os from CD.

Thanks for your response,

Cyb.

 

[Marc Kupper]

My problem is that every so often there seems to be some corruption
happening on one of the disks, the problem has occured on both disks,
on drives D, E,F & G. Small mercy that drive C has not been effected.

I ran into this recently and after much swapping of hardware I realized
the
fault in my case was a newly introduced bug in the NTFS file
compression
code. Google quickly found http://support.microsoft.com/kb/920958/ and
a
hotfix for that issue at http://support.microsoft.com/kb/925308/.

As I had about 2000 corrupted files, though fortunately, no corrupted
directories, I wrote a utility that scans and reports on files that
seem
to be corrupted by that issue.
http://marc.kupper.googlepages.com/scandf

Marc