I think I've got a bum hard drive. Please confirm.

PeregrineFalcon

Basically, any time I either:

- Try to validate the local content for STEAM on Fallout 3
- Run sfc /scannow

I get a crash. First the hard drive light goes solid, then the app in question will freeze up. Then other apps. Then Task Manager. The mouse still moves the whole time, but I have to shut down via the power button.

Things I have tried:
- Deleting the local content for FO3 and recopy it from my network share
- Deleting the local content for FO3 and redownloading it from STEAM
- Running chkdsk /r to check for bad clusters
- Running HD vendor tool (Western Digital Data LifeGuard)
- Running 18 hours of Orthos Prime blend with no crashes
- Swapping SATA cables and ports on the motherboard
- Reinstalling Windows from scratch, using a different disc

Strangely enough nothing else will crash the system. I can play FO3 with the usual Bethesda stability, and all the other Source/STEAM games run fine. It's just these two.

So, well, what the fuck? :?

ronya

If nothing else crashes the system, and the vendor tools and Orthos Prime Blend didn't throw up anything... then just ignore those two bizarre cases until something else pops up?

And keep backups as per usual, etc. etc., but I don't see any reason to suspect an imminent hard drive failure.

PeregrineFalcon

ronya wrote: »

If nothing else crashes the system, and the vendor tools and Orthos Prime Blend didn't throw up anything... then just ignore those two bizarre cases until something else pops up?

And keep backups as per usual, etc. etc., but I don't see any reason to suspect an imminent hard drive failure.

Surprisingly, the guy who burn-tests his rig for 24H with Orthos isn't terribly fond of the "just leave it until it breaks worse" solution, but really since the drive isn't throwing up any fault codes and would just be returned "No Defect Found" it's about all I can do I guess.

ronya

Ahahaha. Yeah, it's dis-satisfactory but if those are the only flaws then I doubt you should do much. :P

I hesitate to ask. Have you tried poking in your Event log to see whether anything interesting happens when it crashes the way you described? While perhaps also getting filemon or whatever sysinternals calls their stuff now to keep writing a dump to disc. Preferably another disc, like a thumbdrive, if you're suspecting HDD fault.

PeregrineFalcon

ronya wrote: »

Ahahaha. Yeah, it's dis-satisfactory but if those are the only flaws then I doubt you should do much. :P

I hesitate to ask. Have you tried poking in your Event log to see whether anything interesting happens when it crashes the way you described? While perhaps also getting filemon or whatever sysinternals calls their stuff now to keep writing a dump to disc. Preferably another disc, like a thumbdrive, if you're suspecting HDD fault.

There's a cockload of:

The driver detected a controller error on \Device\Ide\IdePort2

but there's no time/frequency correlation to the crashes.

I'll give filemon a shot and see if it's accessing one specific file or folder that makes things go boom.

ronya

Well, that particular error can be caused by damn well everything, so that's not helpful.

Regardless, the obvious question now is what's on IDE port 2. Look in Device Manager (Disk Drives > right click every one, one by one, and look at Location. Or flip to Details and look at Physical Device Object Name)

PeregrineFalcon

ronya wrote: »

Well, that particular error can be caused by damn well everything, so that's not helpful.

Regardless, the obvious question now is what's on IDE port 2. Look in Device Manager (Disk Drives > right click every one, one by one, and look at Location. Or flip to Details and look at Physical Device Object Name)

Will have to double-check when I go home at lunch, but I'm pretty sure the answer is "not a damned thing" - it's a one hard drive, one optical system, on logical IDE ports 0 and 1 respectively (physically SATA, but enumeration into IDE is what Windows likes to think.)

Of course if the computer thinks something is on IDE port 2 that's a problem onto itself.

ronya

It's the ghost of your old Pentium's hard drive, come back from the ether to haunt you. :O

PeregrineFalcon

ronya wrote: »

It's the ghost of your old Pentium's hard drive, come back from the ether to haunt you. :O

It did start right after I made plans to sell the old Northwood 2.8C box ...

Perhaps I need to complete the exorcism of the vile system from my domain?

But seriously, I'll check the IDE listings again, and I'm also going to give it a shot when I get a chance with an old IDE hard drive that I'm reasonably sure works.

PeregrineFalcon

Okay, so I'm bringing this back up from the grave for a couple reasons:

1. It's not the hard drive. It passed tests, took a write of zeroes, runs fine in another rig.
2. It's not the cabling. I've tried three now.
3. It's not the software. I'm getting this on both XP and Win7.
4. I can't make new threads.

At this point, I'm thinking it's my motherboard - specifically the drive controller. If it was the PSU, it would be shitting itself in games constantly - which it doesn't do. If it was the CPU or RAM, Orthos would piss itself.

Motherboard is really all that's left. I'm going to try remounting the CPU/memory/board itself in the case tonight as a last-ditch effort. Failing that, I need suggestions for a new LGA775 board that supports 45nm chips. mATX form factor is okay to save money, as long as it's a quality board. I really don't want to build a shiny new rig already - I've had this one less than a year. (And I want i5 to come down in price a bit first, and the ATI DX11 cards to come out.)