Raid 0 Raptor Failure

[Rakewell]

Greetings all,

I have 2 WD74 Raptors in raid 0... the last week or so, I started getting BSOD's that began to occur rarely... however, this morning I booted up my machine 3 times in a row to 3 BSODs, and surprise surprise, the raid array could only detect one raptor, even after CMOS reset. HD failure, right?

I found the out which HD was failing, took it out of my machine and brought it to work and ran the WD Diag Tool in an external enclosure, which found nothing wrong with it. I formatted it, and ran all of the tests again with the same result.

I am using Vista, and I have had these HDs for about 2 1/2 years now with no issues.

So, I guess what I am asking, is do you think that I should RMA the drive, or put it back into my rig and reinstall vista?

Can this sort of thing happen to healthy drives from time to time? Perhaps heat is an issue?

I would love any feedback you folks have.

Cheers,

Rakewell

 

[phatrabt]

I had an experience recently that is somewhat similar to yours. I have a RAID1 array for all my data (not OS), and started to see corruption on large file transfers to the disk. After much t-shooting, I finally isolated it to one of the drives. Removed it and ran the manufacturers diags and the drive came up clean. I then sent the drive in for replacement and they sent me a new one. Installed it and the file corruption disappeared. Moral of the story is that the HD manufacturers diags are great, but not even they can detect all problems. I'd RMA your drive and get a new one.

 

[Rakewell]

Thanks for the reply...

I just remembered something. VISTA did download a RAID update a week ago or so... might that have something to do with this debacle, or in your opinion, could this still be a bad HD?

 

[Smilin]

It ain't Vista :/

The OS you have loaded has nothing to do with whether the controller sees the drive during post or not.

You have a hardware failure of some sort.

Go ahead and RMA it while you're in a down state. When the replacement arrives put the drives on the opposite controllers. That way if you get another failure you'll know pretty conclusively if your controller is the culprit instead of the drive.

 

[RebateMonger]

Failure to see a drive can be caused by cables, the drive itself, or by the RAID controller. If it's intermittant and everything tests good, then it becomes a diagnosis problem. And, of course, loss of a single drive in a RAID 0 array is bad news.

A new client had intermittent reboots on their Server, with a pair of Raptors in RAID 1. They couldn't afford to be down, so we replaced the whole assembly - bought two new Raptors, new cables, and a new 3ware RAID controller. Their previous IT person had put in an $80 Adaptec card, which had terrible ("not extensively tested" according to Adaptec) drivers and was likely the cause of the problem. My point is, it this is a critical array, it might be better to replace everything to get it running reliably. Then test the old drives and keep them as spares for when the new drives fail (which they will eventually).

 

[Fullmetal Chocobo]

What controller are you using?

 

[Roguestar]

People never take into account when building their superfast inevitably raptor level-0 RAIDs:
~RAID 0 fails twice as often as the average hard drive~

 

[Fullmetal Chocobo]

Originally posted by: Roguestar
People never take into account when building their superfast inevitably raptor level-0 RAIDs:
~RAID 0 fails twice as often as the average hard drive~

Some people might not consider it, but you can't say everyone doesn't. I considered it in my build. games & temp go on the RAID 0 array, and that is backed up automatically at intervals.

 

[Roguestar]

Well, ok I was making a blanket statement 😛. Obviously we don't see threads "guys my RAID0 has not failed me ever thanks for the advice" because we only get the "halp @ RAID0 fail" threads, but it seems to be a rather common misconception that people associate RAID with data redudancy/security but don't realise how RAID-0 works 🙂.

 

[Rakewell]

Greetings all...

Spent last night with the bloody thing. It is most definitely the controller on the MOBO, or even the SATA ports as well. I am, or was, using an a8n-sli deluxe. (XP installation wouldn't accept any of my five HD's, raided or not, in any of the sata ports [that board has 8 in all - 4 Raid and 4 non-raid]... I would get to the choose partition section in XP installation, and it would show me "No Disk Inserted" in two different locations on the screen [I can only guess that it is stuck looking for the raid 0 config I used to use]. Every time I hit a key at that point, it would BSOD with no fault reference. Resetting the Bios. does no good, same result every time, no matter what drive or port I used.

It's under warranty, so I guess I could RMA it *(though the crappy Northbridge fan which died the first month I had the MOBO has been replaced with a passive cooler... meaning they may not accept it)...

Should I upgrade to an a8n32-sli deluxe? Anyone know, are there benefits of that board over the a8n-sli deluxe? Newegg has one for 150... not bad, I guess...

...Course I have a monstrous 8800 GTX which barely fits on the damned MOBO. Anyone have an a8n32-sli deluxe that could illuminate?

Thanks,

Rakewell

 

[Smilin]

Originally posted by: Roguestar
People never take into account when building their superfast inevitably raptor level-0 RAIDs:
~RAID 0 fails twice as often as the average hard drive~

Not quite accurate but there is certainly some truth there.

The MTBF of the drives even if lopped in half is well beyond their expected obsolescense. You're also not necessarily doubling the controllers, bus they are sitting on, drivers they are running, software issues they get exposed to etc.

I'm one of those guys running a raptor raid 0 and I've seen 3 external drives and two internal drives die on the rigs in my house during the life of the raid set. It keeps chugging along. I consider myself lucky but not extraordinarily so.

Some single drives have now caught up to it in speed but I've been enjoying the benefits for years before those single drives were available. I would say I'm a raid 0 fan.

I'm no fool though. Even if using Raid 1 it would just be to protect my overall uptime. I have backups to protect my data.

 

[amdskip]

If I was running a raid array I would be running it off a pci card and not the motherboard itself.