Weirder than a salmon eating Frosted Flakes at a red light

[SimplyComplex]

So my computer has lost its mind. Or perhaps my computer problems have made me lose my mind. Not sure which just yet.

I've been working on computers for 16 years now, and I've never seen anything like this. So I'm turning to you for help figuring out what's up.

My motherboard is not wanting to recognize hard drives. Or at least not some hard drives, at some times, in some combinations. All of these checks were done in the bios, not windows(windows also doesn't want to boot half the time with more than one drive connected ATM, but that's another issue entirely).

Here's how things stand ATM.

When I have only one drive connected:
It will recognize my Toshiba drive in any sata port(1-6)
It will recognize my Mushkin ssd only if it's connected to SATA 3
Though it has been recognizing my Hitachi drive on any sata port, now it won't recognize on 1-4. It will regognize it on either 5 or 6.
It will recognize my enterprise Hitachi on port 1-4 but not 5-6

When I have two drives connected:
Toshiba any port(1-6)

Mushkin - nothing not any port 1-6
Hitachi - nothing no port 1-6
Enterprise Hitachi - any port 1-6 will work

When I have 3 drives connected:
Toshiba port 1,2,3 5. Not ports 6 or 4
Mushkin 2,3,5 not 1 4 or 6
Hitachi no port works
Enterprise Hitachi: 5,6, 1-4 do not

4 drives connected:
Toshiba works on all ports
Mushkin work only on 5,6
Hitachi work on 2 and 4, but not 1, 3
Hitachi enterprise won't work on any port


This has been a troubleshooting process of about 3 hours so far. Obviously you may notice one pattern which is ports 5,6 seem to be a bit different than 1-4. They're on a separate part of the motherboard(about 4 inches away from 1-4). You also may notice that the Toshiba works most of the time, and the Hitachi mostly doesn't work. Beyond that, it looks like random noise. I've tried swapping out sata cables. That did nothing. I tried changing the order of the PSU cables. That also did nothing.

And here's the most fun part that I just learned.
You know that whole list up there on how it works when I have X drives connected? All of it is consistent only if those exact 4 drives are present. And it does repeat in that exact fashion. I tested several of those combos after testing other combos, with the same results. But then I tried tossing in one of my old 1tb Seagate drives in place of the Toshiba and guess what? Now the port order is *totally different* even when I put the original drives back in the way they were. With just the toshiba and Mushkin connected, Mushkin now works only on sata 1 and 4. But not 3(like before).

This is an absolute mess. I've never seen anything like it. I don't understand it. And I don't even know what to do about it any more. I tried throwing all of the hard drives in my other computer and they all play nicely there(though it only has 4 sata ports).

I really can't decide if this is a hardware problem with the motherboard or software. I feel like I probably just have to get another motherboard, but I'm concerned if I do, this exact problem will still happen.

Any thoughts on this? Wild speculation welcome. Anything you suggest can't possibly be crazier than what I've seen this morning.

 

[PottedMeat]

try putting pressure with a finger on the controller or bending the board slightly in different areas while booting up, it's probably a intermittent/broken connection somewhere you can't permanently fix ( i.e. a bad solder joint under some bga package or between pcb layers ).

essentally - get a new mobo

 

[deustroop]

I tried throwing all of the hard drives in my other computer and they all play nicely there(though it only has 4 sata ports).

If the drives are properly recognized in another system, its the board that's the problem.

 

[SimplyComplex]

So I decided to actually try to boot into Windows.
The drives that don't show up in my bios, are in Windows now.

Windows is BSODing under high cpu load(over 50%).

I used Startup Recovery

It says ACLs in slui.exe are not proper. And it fixed the problem.

I try to enter my bios.

My bios will not load.

I can however, boot to windows without issue.

I've now tried to disable automatic restart on failure.

I'm not so sure this is a hardware problem. It may be. But I think something is messed up with the UEFI.

This is all started with a random BSOD and Windows freaking out and trying to run 4 different types of tests(2 I've never even seen before).

I think Windows 7 has mucked up the UEFI somehow.

Possibility of a virus is virtually nill, since my system is always in Shadow mode with Shadow Defender(similar to Deep Freeze from Faronics).

I'm going to boot into Windows XP and see if high CPU load causes a crash. I'd bet a fair coin at this point that it doesn't, and XP is going to have zero problems finding the drives.

 

[VirtualLarry]

Perhaps I missed it, but have you mentioned the actual mobo model and CPU, and power supply make and model and age, yet? Please give us that info, it will help trouble-shooting.

 

[DigDog]

seems to me like you have a bad mobo. there's not much you can do beside replacing it.

 

[aigomorla]

i also think your southbridge on your motherboard is probably dying.
OR something on the board is dying.

The only real valid check would be probably to, try all the drives in another board to rule out HDD + PSU.

Or try a different PSU, to rule out board.

If anything is repeatable with new hardware you can rule out that something in the old hardware is bad and needs replacing.

 

[SimplyComplex]

I've already tested with my Seasonic PSU. The problem persisted.

As for what motherboard/cpu it's sort of immaterial but it's an ASRock Z77 Extreme 4, with a 2500k(Sandy Bridge) and 16gb of ram.

I've about 100% ruled out hardware now. I booted into XP and there is no issue. I ran stress tests for 2 straight days and the crashes that were happening in minutes under Windows 7 didn't happen at all in XP.

What seems to be happening, is whenever Windows 7 crashes, the UEFI gets confused on hard drives. I got completely locked out of the bios for a while. But booting into XP fixed it. Upon reboot, I could access the bios again. All of my hard drives are MBR, I don't use any GPT. I stuck it in legacy mode, booted into Windows XP and made sure every drive was showing up. They were. So I rebooted(no power down), went to Windows 7... and all drives are there.

So I decided to do something I've never done before which is go head over to majorgeeks and download some windows repair utilities. I figure at this point I have little to lose. It updates and "fixes" a few dozen things. I rebooted again and everything is still there.

I've been running stress tests on WIndows 7 for over a day and what was crashing it in a few minutes has been running for hours.

I still have no exact idea what the root cause was but it looks like:
?? Windows error BSOD->Auto-reboot->Mucks up UEFI->Can't find HDDs.

The ultimate solution was diabled reboot on crash, booting into XP initially. And then switching to a legacy bios. And then running some random "repair" utilities on Windows 7 to fix the problem.

I guess with a problem this weird I can't expect a repeatable situation. Or a full explanation. It does appear to be resolved now.

 

[Skunk-Works]

I've been following this thread as it's very interesting. I don't know what to make of it. Don't think I ever heard something like this. I have a z270x motherboard and I'm able to write a picture to display on BIOS startup. Well, I recently reinstalled Windows 7 and without all of the drivers installed in Windows, and I think it was the USB 3.0 driver, the BIOS pic never showed up until they were installed. It's weird how the BIOS interacts with Windows. And to top it off I use legacy since I insist on using Truecrypt.