My files change themselves?

Ghosthree3

Apparently my dds files change at random, been getting random "your files do not match the server bla bla" for a while now, it comes and goes and is not always fixed by verify, sometimes it just randomly fixes itself.
It's at the point where I can be IN a server, and when it changes map (to ANY map) I get kicked for a file such as "materials/refinery/tram_wall_clean_01.dds" (note don't have to be actually loading tram), and can't get back into the server, I finally got around to checking the MD5 of this file at some point, and it DOES NOT match the .hashes listed one, but it was FINE less than 10 minutes previous.

Why?

shonan

HDD failing? RAM failing? CPU failing? could be many reasons.

IronHorse

Sounds like file corruption to me.
Steam would only replace the file with the latest, thus your hash would check out just fine, so any change is due to something else, obviously

Ghosthree3

Yeah obviously, I was wondering what it could be, I have been told it could be a bad HD (I've used it for ages, always seemed ok?) but haven't got around to running chkdsk.

shonan

Overclocked your CPU? Could be too big overclock that causes corruption during operations. Also run memtest86+ to check RAM. Chkdsk for drives and http://gsmartcontrol.berlios.de/home/index.php/en/Home to check S.M.A.R.T. status, you can also run a self-test with it that will check whole disk surface and run other test routines built into the drive.

Ghosthree3

I am OC'd at 5.0ghz but it used to do it at 4.4 as well. Also it isn't FREQUENT, it just does happen.

shonan

Are you sure your overclock is stable?

Ghosthree3

Yes.

shonan

Ghosthree3 wrote: »

Yes.

Tested with prime95 and intel burn test?

Ghosthree3

Yes.

ultranewb

Ghosthree3 wrote: »

Yes.

No. You've proven your OC isn't stable. You've an issue that 99% of the time cause by failing hardware, yet you don't see the connection?

Ghosthree3

I don't understand how YOU have decided my OC isn't stable. This issue was happening BEFORE OC, and you say it's caused by it.

EDIT: I'm just so shocked by your statement, who are YOU to tell me my OC isn't stable with absolutely no diagnostics.

IronHorse

shonan wrote: »

Also run memtest86+ to check RAM. Chkdsk for drives and http://gsmartcontrol.berlios.de/home/index.php/en/Home to check S.M.A.R.T. status, you can also run a self-test with it that will check whole disk surface and run other test routines built into the drive.

Did you try any of that?

Ghosthree3

Not yet, the problem with doing tests like that is I need to not have my computer running (well doing anything), that could be a while... I will get around to doing it, for now I plan to check this thread for any ideas until then :P

shonan

Ghosthree3 wrote: »

Not yet, the problem with doing tests like that is I need to not have my computer running (well doing anything), that could be a while... I will get around to doing it, for now I plan to check this thread for any ideas until then :P

You can run gsmartcontrol while using the drives. I'd suggest doing it ASAP to prevent data loss if indeed your HDD is dying.

Ghosthree3

Short self test returned 0 errors, am I doing it right?

Regnareb

You don't and can't prevent data loss by backing up your data when the HDD are dying... that's not what is called a "backup". That's trying to avoid the inevitable.

Regarding the original problem, you did not installed a specific program? A program that sort your files or add metadata to them? I don't know if that could be the source of the problem but... you never know.

shonan

Ghosthree3 wrote: »

Short self test returned 0 errors, am I doing it right?

See basic health check status and more specific smart attributes on the attributes tab, error log is also something you'd want to check. Depending on the drive in question you will have different tests built-into the drive, you should at least run the full extended self test that will check the whole disk surface and I'd recommend running also any other tests you see on the test page.

Ghosthree3

Basic health checks out, I'll upload a screenshot of the attributes, error log is clean. I'll run the full test later.

EDIT: Image here

shonan

You should check out the explanation of raw read error rate, that doesnt look really good. So the basic health check show status as ok? Also your max temperature shows 65, that is way too much for a hard drive.

You may also want to check what head retract cycle count and load/unload cycle count means.

Ghosthree3

I'll check what raw read error rate means BUT the problem is CHANGING files (writing) not reading so I doubt it's related. Every HD shows basic health as PASSED, max temperature doesn't mean frequent, but all the same, what can I do about that, there's a 180mm fan right under it so that's about all I can do.

I'll check what head retract/load/unlock cycle counts are.

shonan

Ghosthree3 wrote: »

I'll check what raw read error rate means BUT the problem is CHANGING files (writing) not reading so I doubt it's related. Every HD shows basic health as PASSED, max temperature doesn't mean frequent, but all the same, what can I do about that, there's a 180mm fan right under it so that's about all I can do.

I'll check what head retract/load/unlock cycle counts are.

A read error will result in corruption, it isnt that simple, since that value is marked as pre-failure it seems very bad, but I cannot say anything as I cannot see the explanation of the attribute in that image.

Ghosthree3

Just so you know, every single one of my HDs (4) have pre-failure on raw read error rate.

shonan

That doesnt help anything as long as you dont look at the explanation of the attribute, different HDDs have different SMART implementations.

Ghosthree3

Where is this "explanation". Is it in the program or google.

shonan

Ghosthree3 wrote: »

Where is this "explanation". Is it in the program or google.

Its in the mouse hover tooltip

Ghosthree3

Updating to say that the extended self-test checked out without error, and here is the description of raw read error rate.

shonan

Ghosthree3 wrote: »

Updating to say that the extended self-test checked out without error, and here is the description of raw read error rate.

Could be that the drive is fine

ScardyBob

I'm curious if you changed your blck to get to 5.0GHz? Also, did you change any other bios settings besides the multiplier and vcore?

Ghosthree3

BLCK is 100 (though the auto tuner puts it at 103), and the following settings were changed.

Multiplier: 50 (duh).
PLL Over voltage: Auto (might have been default not sure).
LLC: Ultra High.
VRM Fixed Freq. Mode: 350.
Phase Control: Extreme.
Duty Control: Extreme.
CPU Voltage: 1.525.

ScardyBob

Ghosthree3 wrote: »

BLCK is 100 (though the auto tuner puts it at 103), and the following settings were changed.

Multiplier: 50 (duh).
PLL Over voltage: Auto (might have been default not sure).
LLC: Ultra High.
VRM Fixed Freq. Mode: 350.
Phase Control: Extreme.
Duty Control: Extreme.
CPU Voltage: 1.525.

I'd uninstall the auto tuner program. I wouldn't be surprised if it overrode some of your settings to mess with the blck leading to corrupted files. I used the Asus Auto Tuner for a while until I realized how atrocious it really was.