Hardware ECC Recovered Count increasing on 3 drives !!

[swamp]

I was copying a large file the other day in linux and got an IO error and lock-up. Very odd for linux as it rarely crashes like that.

I then noticed the Hardware ECC recovered count on the drive increasing constantly as it is used. They were previously zero. I thought it was a bad drive so swapped it for a spare but that showed the same problem, drive was initially zero here but started increasing. Both are Samsung 250G HD250HJ.

I then tried windows 7 running on another drive, a 512G Samsung ECO and that shows the same problem using the free app active disk monitor, increasing Hardware ECC recovered count as it is used !! No bad sectors being relocated or critical faults.


I can't believe I have 3 bad drives and it looks like a m/b controller problem or a psu failing with noise on the lines. I did have a lock-up when burning a CD a few days ago and very rarely the drive fails to spin at boot, maybe 1 in 50 cold starts.

Anyone any ideas. Seems hardware fault, psu is about 5 years old now, OCZ 520
m/b is Asus p5ql pro and is 1.5 years old.

[swamp]

OK, done some more tests.

Borrowed a psu and same problem.
Removed all pci cards leaving just a bare system, m/b, cpu, mem, psu, vid.. same problem
I even re-seated the south-bridge with fresh h/s compound.

This is looking like a Southbridge problem but this is a new fault to me. Anyone else seen drive issues like this. I have the m/b on a good mains filter so not sure how these things just happen ?

Any advise much appreciated.

David

[theycallmebruce]

Interesting problem. I think your best bet is a process of elimination for the hardware, starting with the easiest / most convenient to replace. Do you have any spare RAM, PCI disk controller cards, PSUs etc?

[swamp]

Well, I have tried a replacement psu, all be it a super cheap 400W.
Tested the memory with memtest86+
Moved memory to other slots
Removed all other pci cards and anything else on usb bus, network

Just replaced the vid card and same problem.

Down to either a m/b issue or all 3 samsung drives have started exhibiting the same fault at the same time.


I have been goggling this and there is some mis-information. I have seen statements saying the higher the number the better !! And the common reply is your hard disk is faulty. These values were all originally zero until last week when the problems started.

[Eunos]

I would agree the mb is likely to blame, perhaps a faulty voltage regulator. They can cause some costly damage to other components!

[jfeldt]

Something easy to try is new or different SATA cables. I have had SATA cables give me errors before, even ones that seemed to work fine in the past.