Disk FAT corruption

[dunkers]

It's taken me some time to track this one down and HDSentinel may not be the ultimate culprit, but without HDSentinel there is no problem, and with it running there is.

What happens is that periodically (perhaps every day, perhaps every few days) the file allocation table for my drive F: gets corrupted, requiring chkdsk (or the gui equivalent) to be run in fix mode. I normally notice this when my backup fails due to a volume snapshot failing to be created. The error returned is:

ERROR: COM call "m_pVssObject->BackupComplete(&pAsync)" failed.
- Returned HRESULT = 0x80042301
- Error text: VSS_E_BAD_STATE

After I run chkdsk in fix mode the backup will perform flawlessly.

The strange thing is that this only affects drive F:. My setup is a pair of 1TB drives in a RAID1 (mirror) array. The array is then partitioned into a number of primary and logical drives, 9 in total. None of the other drives is affected.

The only thing that regularly uses drive F: is TrackMania, which I play quite a lot. But playing it with HDSentinel running doesn't appear to cause a problem. I think it must be the kind of problem where two events occur at the same time, and much of the time the events miss each other. Although I have shown that HDSentinel is involved (by not running it for lengthy periods during which there are no problems, then running it and seeing problems), I haven't performed the same test with Trackmania not run. I would simply die of boredom

Is there some way I can close in on this problem to determine the cause and, hopefully, a fix?

[hdsentinel]

Thanks for the information!

From the details you sent, I can confirm that It is not really related to Hard Disk Sentinel.
Hard Disk Sentinel communicates with the physical hard disks in general, so if it would be the source of the issues, other logical drive(s) could be also affected on the same physical drives (the RAID set). Hard Disk Sentinel can't cause issues on a logical drive as it does nothing (except displaying capacity and free space) of any logical drive, for example your F: drive.

Personally I'd feel the issue may be related to a hard disk controller driver which may be required to be updated or a minor issue with one hard disk in the position of the logical drive F:
I'd check if there is a driver update available for the hard disk controller and/or use the safe tests in Hard Disk Sentinel (Disk menu -> Short self test, Extended self test, Surface test -> Read test) to check the stability of the hard disk.

If you can use Report -> Send test report to developer option, I can investigate the current status of the drives, the hard disk controller drivers and advise based on the current situation.

[dunkers]

Thanks for the quick response. I just ran off a report and sent it to you.

I don't think it's the disks, which are both seriously new. I guess the RAID driver might be a good match since it's that which arbitrates accesses. It's the problem always occuring on the F: drive which puzzles me - that has to be due to normal software.

I've just thought of some more tests, so I'll try those and report back

[hdsentinel]

Thanks for the disk report!

Yes, it seems the drives are perfect, as reported by the software, just the driver of the RAID controller (the Intel chipset driver) seems a bit old.
I'd recommend to update to the latest working version (10.5) which is available directly from Support -> Driver Zone on www.hdsentinel.com

[Ernst@at]

A possible influence could be:
HDS submits DEVICE IDENTIFY cmd to Samung HDDs while they are working down other requests, so some Samsung HDDs "forgets" sometimes to write contents to surface. There is a firmware update available for their affected models.

[hdsentinel]

Not really....

HDS checks for drive type to prevent such issue. Also the user had Seagate + WDC drives according the report he sent for investigating the issue.

[dunkers]

I don't think it could be at such a low level because this is a RAID pair, and if they get out of step I would be looking at many hours of lag whilst they re-sync. Also, I can't see how that would only affect a specific logical drive every time.

Sorry I've not been back to update this thread. I haven't forgotten it at all - it just takes a long time to make a small change and then wait a week or so to ensure that's done something (or not). So, I've tried alternative disk monitors and the problem occured with those as well (albeit not quite so often). So it's not HDSentinel that's at fault. I updated the Intel drivers and that was a massive improvement, but not a complete fix.

By one of those strange quirks of fate, this morning I plugged in a newly purchased USB stick and, on ejecting, saw it marked as "AutoRUN/partition". WTF!?! This PC is supposed to not allow anyt sort of auto-anything, so I was somewhat worried about auto-installed malware and took a deeper look at the system. Every drive had snapman.sys listed as a driver, and this turns out to be an Acronis thing. The last time I had to do a disaster recovery was because I used Acronis Disk Director to re-size a partition. You can see where this is going...

So, I have now got rid of that and am optimistic that my particular problem is now resolved.

[hdsentinel]

Thanks, glad to hear the issue seems resolved !

Yes, Intel drivers are not really perfect ... Even newer version of Intel drivers have some issues and we constantly need to make new "fixes" for these buggy drivers

Anyway, if possible, please use Report -> Send test report to developer now, after the upgrade of drivers. This way I can check and compare with the previous status received. Also it may help if any issue will appear (hopefully there will be none).

[dunkers]

A report is winging it's way to you as I type

[hdsentinel]

Thanks!!
Yes, now things seem perfect!! The current Intel driver (10.5) is far the best version.

[celavey]

Will it depend on the hard drive you are using? What happened on this then?

[dunkers]

What happened on this then?

It didn't get fixed, but it remained at a more or less low level. Until last week when I discovered two things:

1. Unreal Tournament provokes it.

2. The reason I stopped playing UT was because of the disk corruption.

3. I have a short memory.

However, over the weekend my video drivers (ATI/AMD) were updated, and despite getting stuck into new maps on UT I haven't had a repeat yet.

[dunkers]

Sorry to ressurect this one, but I have finally resolved this and thought I should finish this off.

Some time back I used a partition manager to adjust a partition that was running out of space. The tool managed to lose four partitions, and instead of restoring from backup I chose to recover the partitions by searching the hard disk. It was much faster than a restore that way, and everything got back to how it was.

Except... the F: drive was one of the recovered partitions. Last week I deleted the partition properly and then recreated it and restored from backup, and it's been fine ever since. So it seems that the original was recovered well enough for the partition to be used but with a slight flaw that sometimes caused corruption (but I don't know the details of how that worked).

So, the moral of the story is twofold:

1/ Never ever use Acronis Disk Director on a system you might want to use afterwards. Actually, I would say the 'disk director' part of that is superfluous.

2/ MiniTools Partition Wizard is a life saver. Earlier today, for instance, I resized my C: driver whilst it was live and running. Hard to think of a trickier task to pull off than that.