OK Geeks, answer this one....

[MilDoc]

I have a Dell XPS 600 system with built in RAID support (set to RAID 1) via an nForce4 chip set (nVidia).

Yesterday when booting I got a "deranged" error on BOTH drives.

I went into Setup and turned off RAID support for both drives (C & D) and BOTH check out using every drive checker I have.

That suggests to me that the nVidia SATA-RAID onboard chip set has failed.

I have spent hours trying to find a diagnostic for this chip set, to no avail. Anyone know a source?

Well if not I still have almost a year of warranty support from Dell, but using it is a PITA (Had to do so when the DVD drive failed. I KNEW what was wrong, and spent 2 hours on the phone until the tech agreed with me and sent a new drive).

So,

1) agree with my diagnosis, failed chip set? If not - what, and how to diagnose?
2) anyone know of a test program for the nVidia4 SATA-RAID chip set on the motherboard?

I have REALLY grown to hate on the MB chips. IMHO, better to have separate cards. But that is just my opinion, and I could be wrong.

[Bill Stock]

Sorry, I've been deranged for years, so I can't help.

[crokett]

Easiest way to diagnose is go buy another SATA controller. Doesn't even have to be RAID assuming you were doing a simple mirror. Put the controller in, if the drives work it is the onboard controller. Then call Dell and get the onboard fixed and return the one you bought. Just make sure wherever you got it from doesn't have a restocking fee.

[cgallery]

I think they were "degraded" rather than "deranged."

Anyway, this isn't atypical of RAID configurations. The most likely culprit is one of the drives, not the chipset. It is possible that it is the chipset, but drives are substantially less reliable than chipsets, and are prone to intermittent failures. A hard drive that is having a hard time (slow) writing a sector, for example, can cause the RAID firmware/drivers to degrade the RAID volume (running off the drive that WASN'T slow).

If you have any utilities that will list the SMART tables for the drives, comparing the two may give you a clue. Drives with "reallocated sectors" or "Ultra ATA CRC errors" may be bad, or have a bad cable (with Ultra ATA CRC errors also sometimes a sign of a bad motherboard host).

That is, if you look at the SMART tables on the two drives and compare them to one another, the answer may become obvious. A bad drive will often stand out.

If the tables look nearly identical, then I'd check the motherboard's firmware (see if it can be upgraded), and upgrade the nVidia drivers, too.

It also wouldn't hurt to check your motherboard for leaky caps. They can cause all sorts of strange things.

But I have probably 100 machines with nVidia RAID in place in hospitals and clinics all over the country. It works pretty well. We see a degradation about every other month. Sometimes we just rebuild the array and never have a problem on that machine again.

[Kristofor]

I KNEW what was wrong, and spent 2 hours on the phone until the tech agreed with me and sent a new drive).

Cgallery gave some good advice on your current issues, so I'll address the DVD problem above.

In most call centers (my full time job for 3 years in college), it doesn't matter if the tech is almost certain he knows what the problem is. The company (at least the 2 I worked for) would rather pay some additional labor than ship parts out when it's not a hardware problems . As a result there are some standard diagnostics processes that must be followed before the tech can send out a part (large business accounts often have a different standard of service.)

Knowing this is how they work, you need to know how to "properly" describe the problem to the tech. Here's an example.

<intro BS with serial numbers, names, etc>
You: "My DVD drive died I need a replacement"

Tech: "Okay, we just need to check a couple things. This should only take a few minutes <lie>. Can you start by..."

You: "Oh, you don't understand, it started making a buzzing or grinding noise, and then smoke started wafting out of the drive. I took the side off the case and could see that it was coming from the DVD drive and not anywhere else inside. I disconnected the drive and took it out of the computer as my neighbor recommended (he works for the fire department). The computer seems to work now but I don't have a D: drive any more, do you think that has something to do with it?"

Tech: "Uh, probably... Okay, we'll have a new drive to you on Wednesday"

You need to give the tech the tools/freedom to help you. Fire/smoke will almost always bounce you out of any diagnostic flow chart... Of course, if you don't have the spare hardware to test with, and aren't sure that the part is really bad, then this is not a very efficient or ethical approach.

[MilDoc]

I think they were "degraded" rather than "deranged."

Anyway, this isn't atypical of RAID configurations. The most likely culprit is one of the drives, not the chipset. It is possible that it is the chipset, but drives are substantially less reliable than chipsets, and are prone to intermittent failures. A hard drive that is having a hard time (slow) writing a sector, for example, can cause the RAID firmware/drivers to degrade the RAID volume (running off the drive that WASN'T slow).

If you have any utilities that will list the SMART tables for the drives, comparing the two may give you a clue. Drives with "reallocated sectors" or "Ultra ATA CRC errors" may be bad, or have a bad cable (with Ultra ATA CRC errors also sometimes a sign of a bad motherboard host).

That is, if you look at the SMART tables on the two drives and compare them to one another, the answer may become obvious. A bad drive will often stand out.

If the tables look nearly identical, then I'd check the motherboard's firmware (see if it can be upgraded), and upgrade the nVidia drivers, too.

It also wouldn't hurt to check your motherboard for leaky caps. They can cause all sorts of strange things.

But I have probably 100 machines with nVidia RAID in place in hospitals and clinics all over the country. It works pretty well. We see a degradation about every other month. Sometimes we just rebuild the array and never have a problem on that machine again.

OK. Checked for leaking caps. All look fine. Also made sure all boards, cables, etc, were firmly seated.

Ran SeaTools for DOS (no Windoze interference that way - direct boot from CD) on both drives. SMART tables same. Drives both passed the "Long" test, which also means NO bad sectors, firmware checks out, etc.

Dell diagnostics checks the cables. All AOK.

And I have the latest Dell MB firmware update and the latest nVidia drivers. That's something I look for every 3 months, on schedule.

Now, when I go into the RAID utility at setup, and push "R" for Rebuild, NOTHING happens.

Still seems to me to be the nForce3 chipset, but I have yet to find any program that can diagnose that chipset.

Anything else to do?

[cgallery]

Now, when I go into the RAID utility at setup, and push "R" for Rebuild, NOTHING happens.

Still seems to me to be the nForce3 chipset, but I have yet to find any program that can diagnose that chipset.

Anything else to do?

Oh, actually, I have found that the "R" doesn't work in some instances (like when the BIOS can't figure out which drive has the live data on it).

So I'd boot into Windows, go to: Start/Programs/All Programs/nVidia/MediaShield.

Within MediaShield you can tell the RAID to rebuild.

You may be required to delete the array, which will essentially tell MediaShield to break the array. But then it will take the "new" spare drive and re-add it to the mirror and start rebuilding.

If you get stuck just shoot me a PM and I'll give you my cell #. I can either walk you through it or I can do it for you w/ our remote-control app.

[MilDoc]

Oh, actually, I have found that the "R" doesn't work in some instances (like when the BIOS can't figure out which drive has the live data on it).

So I'd boot into Windows, go to: Start/Programs/All Programs/nVidia/MediaShield.

Within MediaShield you can tell the RAID to rebuild.

You may be required to delete the array, which will essentially tell MediaShield to break the array. But then it will take the "new" spare drive and re-add it to the mirror and start rebuilding.

If you get stuck just shoot me a PM and I'll give you my cell #. I can either walk you through it or I can do it for you w/ our remote-control app.

Will that result in any data loss? That's what I'm afraid of. Never had a RAID array before this one.

[cgallery]

Will that result in any data loss? That's what I'm afraid of. Never had a RAID array before this one.

The post by bnborg on Aug. 26, 2008, covers it:

http://forums.nvidia.com/lofiversion...hp?t70473.html

Decide which of the two drives to call good. Usually it will be the one mounted on the old array drive letter.
Go to Disk Manager or Device Manager and find out which one it is physically (location) by controller, etc.

Reboot and go to the NVidia Raid Manager by pressing F10 (or whatever key is indicated) at the right time. You can also use MediaShield, in Windows, if it will run for you.

Delete, yes Delete, the extra "Degraded" array.
Highlight the remaining "Degraded" array.
Select Rebuild, then add the other drive that is now available.

The reason I suggest doing this in MediaShield instead of the BIOS is that once you're booted into Windows and running MediaShield, it is pretty difficult to do anything to the disk configuration that will cause you to lose data. That drive you booted from will be your source drive and is somewhat protected by the fact that Windows is running from it, and the drive you delete from the array will become the target.

Many times I don't even have to select "Rebuild." Depending on the version of the RAID drivers, sometimes once you delete the extra degraded array, the drivers will automatically start rebuilding on it again.

Kinda scary, but go ahead and do it.