Major HDD and Graphics Corruption. Help!

[WLC-1989]

Hey everyone, I'm afraid this is going to be a bit of a long one...

Might I first reference a thread I made recently:

http://forums.hexus.net/operating-sy...reen-tabs.html

Since I made that thread, I installed Battlefield 2. Evidently this put my computer through more strain than most games would, and I began to experience a number of artifacts in-game, as well as in Windows after playing the game. Some of these post-game artifacts and graphical oddities were similar in nature to the ones I described in the aforementioned thread.

My core PC specification is as follows:

DFI LANParty LT X48-T2R
Intel Core 2 E6600 2.4ghz
HIS ATi X1900XT 512mb
4x1GB GeiL 4-4-4-12 800mhz DDR2
620w Enermax modular PSU
Vista Ultimate SP1
8 hard drives of differing size (including the ones discussed below).

Anyway, a little while ago I purchased two new hard drives, 1000gb and 750gb Samsung F1s. Both had been running without a problem, but recently I've been experiencing significant slow-downs when trying to access my 750gb drive. I use it for storing media files which are generally around 100mb to 1gb each.

These slow-downs often present themselves as Vista just loading the folder (as is so often the case with large folders in Vista). However, about a week ago, this loading continued for about 2 minutes, and then ended with an error stating the folder could no longer be found. I returned to the directory this folder was in, only to find it had completely disappeared.

Obviously concerned that I had lost data, I ran a recovery program, and found to my surprise this whole folder in perfect condition. It seemed it had been orphaned: It's file allocation 'index' had been lost.

A few days ago, the same thing happened again, and I was presented with a 'this folder is corrupt' error. I restarted, and CHKDSK ran. It found that another bunch of files had been orphaned, and recreated the indexes. This has happened every time I try to browse this drive. To my surprise, every time CHKDSK restores the index, the files themselves are in perfect (uncorrupted) condition.

In essence, something seems to be corrupting the indexes.

To add to the mystery, yesterday, my main boot drive (a 200gb Samsung) began giving corruption errors. Almost every program lost vital files and was unlaunchable, with core Windows programs and services crashing. I restarted and ran a full CHKDSK test on it, and found that no less than 180,000 files had been orpahaned. The process of reconstructing it took in excess of 40 minutes. The same thing happened again within about 30 minutes of being back in Windows.

I have always had a full antivirus and firewall up (presently NOD32), and have never experienced such strange index-loss problems before. Furthermore, I reinstall Windows regularly, with the safest (often latest) drivers available. My first suspicion was the RAM; I ran Memtest86+ overnight, and it found no errors in any of the RAM I use.

I can therefore only conclude that one of the following is happening:

1) 2 different models of Samsung hard drive, bought a year apart, are failing simultaneously from apparently the same (strange!) problem. In addition to this, my graphics card is failing for some other reason.

2) My PSU is failing, despite it being from a good brand and having no previous faults, and without affecting any of my other drives.

Neither of these explanations really seem to account for all of the problems I'm experiencing. I am aware that I could be experiencing two different hardware failures at the same time, but I'm not sure what exactly could be doing it. Over a month ago, I changed to the new DFI motherboard, but I did not experience these problems directly following this, and so have largely discounted it as a cause.

Can anyone recommend a course of action and/or specific ideas as to what is going on? I'm completely stuck.

I appreciate you reading this essay of a problem, and any advice you can give.

[WLC-1989]

Update:

I re-ran the extended CHKDSK program on the 750gb drive. It hanged at part 4 of 5, and concluded by stating:

"The disk does not have enough space to replace bad clusters".

I'll update this thread further in the morning; I'm currently trying to backup as much as I can, though it seems certain things like 'FOLDER.jpg' files are getting corrupted.

[peterb]

First thought is that it could be system memory - download memtest 86+ and check out the memory.

Second thought is that the PSU is on its way out and cousing soome disk corruption.

It is unlikely that two disks of different sizes/production runs would fail spontaneusly in the same way at the same time - without an external cause.

[WLC-1989]

Thanks for the reply; as I previously mentioned, I ran Memtest overnight with no errors found.

I agree that the PSU seems the best bet, but I don't see why it would affect only two (quite different) hard drives, and not the others?

Thanks for any help you can give.

[Jay]

I had 2 HIS X1900XT cards that where faulty in the way that you mention. I gave up with HIS in the end and got a Sapphire card.

I hope you didn't get the card from OcUK

[WLC-1989]

Unfortunately I did purchase my X1900XT from OcUK...

Still, I think it only came with a 1 year warranty, in which case I'm out of time already.

At the moment, then, it's looking like I have a faulty graphics card, and 2 dead hard drives. Talk about bad luck...

[Jay]

yes, thats where my HIS cards came from as well. They just could not supply one that didn't die after 1 day.

[WLC-1989]

Ok, another update. I reformatted the 750gb drive, and copied the salvaged files (as well as some other stuff) over so as to find out if it would corrupt again. So far, I've had no problems.

Does anyone know if another corruption is likely, or given that two drives corrupted at once, does this seem like just 'one of those things'?

[noveltylamp]

Use HDTune or Everest to examine the SMART tables for the HDD. If there are any reallocated sectors then they have sustained physical damage and will corrupt again.

I think Speedfan has a SMART diagnostic too.

[WLC-1989]

At the moment, it says:

Threshold: 10
Value: 100
Worst: 100
Data: 0

This is just after I refortmated. It does however have a '1' under the data on 'error rate'. I'm not sure if that's of any significance, though I'm not sure how to interpret any of this.

[peterb]

Use HDTune or Everest to examine the SMART tables for the HDD. If there are any reallocated sectors then they have sustained physical damage and will corrupt again.

I think Speedfan has a SMART diagnostic too.

Not strictly true - drives do have defective sectors at manufacture, and these are mapped out and spare sectors - provided for that purpose - mapped in. During the life of the drive, other sectors can suffer soft errors and when they reach a threshold, the data is copied to another sector - and that one mapped out. This is all done by teh drive electronics and is transparent to the user - but reported by interrogating the smart data. Hard errors caused by physical damage are more serious - and more likely top be terminal for the disk as it can cause head damage.

That said, a large number of sector errors or seek errors may be indicative of impending failure opf the disk - although the when is much harder to predict.

This link

http://research.google.com/archive/disk_failures.pdf

gives details of a research paper by some Google emplotyees (Google use thousands of disks!) and make interesting reading - but be warned - it is an academic paper and isn't light reading!