The file system structure on the disk is corrupt and unusable. Please run the chkdsk

[mattx]

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

 

Hmmmm, lots of nice messages regarding this and now getting reports of users reporting corrupt documents etc. Anyone got any ideas on how to run a chkdsk on a iscsi drive ?

 

Will a chkdsk /x mean users losing access to the mount ?

Edited November 19, 2010 by mattx

[mattx]

This is a nightmare - I've got corrupted files all over the place on a 900GB iscsi disk - I've forced a /x chkdsk thus upsetting quite a few users - but I have my doubts this will have sorted this. Has anyone had experience of this sort of thing ?

I've checked the logs on our SAN and I don't have a drive failiing.

[mattx]

Well I have to say I've never seen or experienced anything quite like it - I've managed to stop the corruption but of course there are plenty of files on our san with 0 bytes of value and some which are corrupt. Thank god for backup tapes - will be using the full backups from last weekend to recover data that users report. I would love to know why all of a sudden I had an influx of corrupt data on our SAN. This all stems from a power outage we had Tuesday afternoon, on checking the UPS's did their job, nothing went down and power came backup after about 10 mins. So I am now left wondering why this has now caused a problem - specially with the file corruption. Any power surge should be taken care of by the UPS's ? Or am I wrong. I would appreciate anyone's thoughts. Edited November 19, 2010 by mattx

[DAZZD88]

I know this might ounddaft but have you tried a basic CHKDSK H: /R?

 

We had to run this on our RAID5 set upon our server when our server room got really hot and the disks in our server died. Luckily one was only temporarily out of action which allowed us to rebuild the RAID array and run CHKDSK to repair some damaged files. All in all I think we only lost a few minor files, if that.

[bossman]

@mattx:

 

Could it have been the controller unit on the drive or iscsi controller card which may have been damaged due to a large power spike even though the UPS didn't suspect anything untoward this may have been the problem.

 

How is the drive reacting now?

[LiamH]

Are your switchs/network gear attached to the UPS?

[mattx]

@mattx:

 

Could it have been the controller unit on the drive or iscsi controller card which may have been damaged due to a large power spike even though the UPS didn't suspect anything untoward this may have been the problem.

 

How is the drive reacting now?

 

After running a chkdsk /x the wads of error messages have stopped - thank god. I have recovered quite a few corrupted files off our backups but I know there will be more. It's running ok now as far as I can tell but obviously the file system on the LUN is damaged. Question is - do I leave it as is [ take a full back up this weekend ] and see how I go or start again ?

I've double checked the SAN and there are no reports of any failures, no reports of anything else on the main server either. I'm hoping I stopped this before it became too serious.

[mattx]

Are your switchs/network gear attached to the UPS?

 

Yes - all plugged into the UPS.

[mattx]

I know this might ounddaft but have you tried a basic CHKDSK H: /R?

 

We had to run this on our RAID5 set upon our server when our server room got really hot and the disks in our server died. Luckily one was only temporarily out of action which allowed us to rebuild the RAID array and run CHKDSK to repair some damaged files. All in all I think we only lost a few minor files, if that.

 

Did not force use the /R - just did a basic chkdsk with a /x to force a dismount. I did not want it to take an age which it may do as it's 900GB in size.

[mattx]

On monday I'm going to dump the entire contents of the tape which I have been recovering corrupted files to a spare partition on our server - that will save me having to load and re-load the tape - out of a load of data already recovered I have only been unable to get back 2 files which were backed up to my daily incremental tape with 0 bytes....Aaagggh.

[DMcCoy]

I assume only one server is attached to the iscsi LUN?

[mattx]

I assume only one server is attached to the iscsi LUN?

 

Yes and data on the server partitions [ Not the LUN ] are fine. It's just data on the LUN. After running the chkdsk /x this morning it seems I have stopped the influx of 'The file system structure on the disk is corrupt and unusable.' messages etc.

It's just a case now of recovering data when it gets reported to me from our backups.

[mattx]

Right, just ran a chkdsk without any switches before I go and fight the Friday evening traffic on my bike and this is what was reported:

 

943714295 KB total disk space.

86742528 KB in 262440 files.

125272 KB in 93045 indexes.

0 KB in bad sectors.

451591 KB in use by the system.

65536 KB occupied by the log file.

856394904 KB available on disk.

 

4096 bytes in each allocation unit.

235928573 total allocation units on disk.

214098726 allocation units available on disk.

 

I shall have my fingers crossed all weekend now. But I think I shall have to start looking at an alternative when it comes to it running off two UPS's [ one on each PSU of the SAN ] - anyone got any suggestions ?