DMA CRC Error Count is very high. Sata controller failing?

[nrn9520]

HD Tune is showing very high numbers of Ultra DMA CRC ERRor Count for both drives running on my motherboards SATA controller. Over 63,000 for the system drive! I have a 320GB WD, and a 64GB OCZ Agility SSD as the boot drive. Here are the Health screenshots:

Agility


WD


HD Tune says that it could be caused by a damaged cable, but i have changed both cables. I have also moved them from ports 0+1 to 3+4. This has had no effect.
Is this a sign of the SATA controller failing? The system has been working fine, except that games tend to crash for some reason (even though CPU and Ram stress tests are always stable). Any info about the situation would be appreciated.

Edit: Motherboard is Gigabyte P35-DS3L

 

[almartin]

It could be because when your game crashes it is logging these errors. You could try and boot to the dos prompt and run CHKDSK to check your drives and let it repair any bad sectors if there are any before you load windows. Also clear your log files and get a fresh update on your errors.

 

[sminlal]

These indicate problems with the transfer of data between the host and the disk. They can't be caused by software, and they don't indicate a problem with the disk itself. See: http://en.wikipedia.org/wiki/S.M.A.R.T.#Known_ATA_S.M.A.R.T._attributes

The first thing I'd do is to reseat all of the drive cables. If the numbers continue to climb then I'd try using different cables.

Edit: Ah, I see you've already done that. Given that it's happening in multiple drives it seems pretty unlikely to be a drive fault. But could still be some systemic problem such as an overtaxed power supply.

 

[nrn9520]

Yes, I assumed like you said that multiple drives eliminated them as the problem. The first thing is did was change cables, and also change to different connectors on the motherboard.

My power supply should be having no load problems unless it is failing. I have a Corsair HX620 powering E8400, GTX 560Ti, 4GB Ram, X-fi, and PCI Raid controller with 2 hard drives in addition to the ones showing problems. I don't think that should be too heavy a load, correct me if I'm wrong on that.

The process of elimination still leads me to believe the ICH9 is having issues, possibly an early sign of larger problems with the motherboard as it begins to fail. Is this a sensible conclusion, or one that isn't even possible?

 

[sminlal]

Anything's possible, but failure of the chipset itself seems unlikely to me. Of course it could be something that for all intents and purposes is the same, such as bad traces or solder joints on the motherboard.

Did you say you have two OTHER drives that aren't showing problems? Are they plugged into ports on the same controller? A logical thing to try would be to swap the connections of the working vs. problem drives to see if the problem moves with them or not.

 

[nostradamus99]

Anything's possible, but failure of the chipset itself seems unlikely to me. Of course it could be something that for all intents and purposes is the same, such as bad traces or solder joints on the motherboard.

Did you say you have two OTHER drives that aren't showing problems? Are they plugged into ports on the same controller? A logical thing to try would be to swap the connections of the working vs. problem drives to see if the problem moves with them or not.

Hi sminlal,
No intent to hijack the thread. I sent you pm last week regarding this thread: http://www.tomshardware.co.uk/forum/267900-14-optical-external-back If you have a minute to spare it would benefit a lot of people..
Thanks again!

 

[nrn9520]

No, they are not on the same controller. They are in a RAID 0 array on a PCI SiI raid controller (motherboard does not support RAID).

Technically I do not know if they are showing the same error or not, as I can't see the drives health status because of running in RAID.

 

[sminlal]

Not that this will solve your problem, but it might avoid it - if you have spare ports on the RAID controller you could move the drives there as JBOD (individual, non-RAID volumes)...

 

[sminlal]

Hi sminlal,
No intent to hijack the thread. I sent you pm last week regarding this thread: http://www.tomshardware.co.uk/forum/267900-14-optical-external-back If you have a minute to spare it would benefit a lot of people..
Thanks again!

Your link is to the Tom's UK site, which I don't seem to have a sign-on for. Your question in the thread was:

Hi sminlal,
Can you please explain how and what util you are using to checksum your data? would be very interested..!
Many thanks in advance..!

I actually wrote my own utility, but there are plenty of free ones available if you Google "Checksum Utility".

What I do is run the utility to create a file containing checksums of all of the files on the disk. Then later on I can re-run the utility to verify all the checksums. If any of the bits in the files have changed between the time the checksum file was created and when the verify was done, the checksum utility will report a checksum mismatch.

Obviously if you change any files on the drive then you'd have to run the utility again to update the checksum file. I mostly use this for folders that contain "archive" files which never change (such as photos, music, etc) or backup files.

 

[nrn9520]

Not that this will solve your problem, but it might avoid it - if you have spare ports on the RAID controller you could move the drives there as JBOD (individual, non-RAID volumes)...

It's only a 2-port controller, so I will not be able to do that.

I guess I will just have to keep frequent backups and hope that it keeps working normally as it has been, until I get a new SB system built.

Thanks for your advice on this

 

[sminlal]

Are the CRC numbers actually still increasing over time? If not, you may actually have solved the problem when you fiddled with the cables.

 

[nrn9520]

Yes they are. The OS drive has about 1000 more now, so no luck with that.

 

[sminlal]

I wouldn't be very happy with that. I might go so far as to buy a cheap SATA to PCIe adapter card and try using it instead of the motherboard ports to see if it solves the problem...

 

[nrn9520]

That is a good thought. I believe my only available PCIe slot is inaccessible due to my GPU cooler though. (I'll have to check when I get a chance.) Would having both drives on an old PCI card limit my SSD's throughput?

It's advertised specs are:
Read: Up to 230 MB/s
Write: Up to 135 MB/s

 

[sminlal]

It would probably limit the throughput with a cheap 1-lane PCIe card, but only if you're pushing I/O at the maximum rate to both drives at once. But SSDs perform well mostly because of their fast access times, and that's not affected by a transfer rate bottleneck.

 

[nostradamus99]

Your link is to the Tom's UK site, which I don't seem to have a sign-on for. Your question in the thread was:


I actually wrote my own utility, but there are plenty of free ones available if you Google "Checksum Utility".

What I do is run the utility to create a file containing checksums of all of the files on the disk. Then later on I can re-run the utility to verify all the checksums. If any of the bits in the files have changed between the time the checksum file was created and when the verify was done, the checksum utility will report a checksum mismatch.

Obviously if you change any files on the drive then you'd have to run the utility again to update the checksum file. I mostly use this for folders that contain "archive" files which never change (such as photos, music, etc) or backup files.

Thanks sminlal.
I've found these 2 utils for now I'll see if it does the job..

fastsum
http://www.fastsum.com/

and

Advanced CheckSum Verifier
http://www.irnis.net/soft/acsv/
Thanks again for your help!

 

[asasa]

Major bump. I also have this problem. My computer freezes often, and upon bootup will not recognize my drives.

Interestingly enough, I have an OCZ Agility 64gb as well, alongside a WD 640gb.

Other hardware: ASUS P9X79 PRO / i7 3820 / 4x4gb G.Skill 1666
Win 7 64bit

I've reformatted and reinstalled once already, as the old install that ran flawlessly for over a year could not be repaired.

I've used multiple cables, tried using the different SATA controllers on the mobo, etc. Any ideas?