Drive disappeared from RAID0 array

[zefyx]

I came home earlier to try to extract a archive on my 4.5TB RAID0 array (powered by Intel's ICH9 chipset) to be greeted by WinRAR saying the archive couldn't be extracted as the disk might be full. After trying a few times I decided to reboot the computer, which I tried 3 times & received the same message. I tried then tried switching in my bios from RAID to IDE to run Seagate's SeaTools to test the drives, SeaTools would not run & gave me & error which wouldn't allow it to run. I then rebooted the PC & switched my drives back to RAID. PC shut off & turned back on. At this point, Intel's RAID boot-up told me my array named 4.5TB had failed & that one of my disks was no longer listed as a member of the array. I'm basically panicking at this point. Windows is taking quite awhile to boot, I'm unable to test the drive that was kicked from the array in SeaTools (SeaTools says Test Unavailable). I've tried testing in DOS, apparently SeaTools doesn't support GPT. DiskMgmt takes along time to connect to logical disk manager.

Some information:
Windows XP Professional 64-bit Edition
(3) 1.5TB Seagate HDDs
GPT formatted disks
Intel ICH9 RAID0
(3) drives no older than 1 month
(2) drives with CC1H firmware
(1) drive with CC71
Drive that is missing does NOT make any strange noises

The worst part of all of this is I have another 1.5TB drive that is supposed to be delivered tomorrow, I ordered this drive so that I may convert my RAID0 array to RAID5 for redundancy.

What do I do? "Some" data was written to the array which presumably was during the time one of the disks left the array, however the data that might have been written at that time is unimportant to me, I simply don't want to loose the other 3TB that were on these disks. Please help!

Also, I'll say this so the flamers don't have too. I know it was in RAID0, I know I should backup if I'm worried about data. My only excuse is I had & planned on having this RAID0 array for 2-3 weeks until I could pickup another 1.5TB to convert this array into a 4TB RAID5 array. Please help if you need any other information, please let me know.

Here are screenshots:

DiskMgmt.msc screenshot: http://i31.tinypic.com/6px14z.jpg
Intel Matrix Storage Controller screenshot: http://i29.tinypic.com/ftkiz4.jpg
Seagate SeaTools screenshot: http://i31.tinypic.com/s3n6dv.jpg
HDTune screenshot: http://i26.tinypic.com/15i4zft.jpg

 

[RebateMonger]

I am NOT an expert at RAID array recovery. I've only had one client that was willing to pay for hardware recovery of a failed striped data array. That recovery, a Windows software-based RAID 5 array with three SCSI drives and an apparent dual-drive failure, cost about $25K.

This is potentially a very difficult problem. You are using data striping along with an uncommon partition type (GPT). That partcular drive (Seagate 1.5 TB) has a reputation for early failure. The reason it's taking so long to boot is that one of the drives isn't responding correctly.

With a RAID0 array, you don't lose 1/3 of the data when a drive fails. You lose ALL of the data. Unless you can get the failed drive to deliver some sort of data stream, software data recovery tools won't work. I guess that's where the effort should be. Getting an image from that drive. You can then work with that image, along with the other two drives to do recovery from the array.

Unless you can wiggle connectors and get the drive running again, this is likely a job for experts. Sorry.

 

[zefyx]

Thank you for the reply.

Unless you can wiggle connectors and get the drive running again, this is likely a job for experts. Sorry.

The drive is running now, and never stopped running to my knowledge. Windows see's the drive, however it refuses to go back into the array for some reason.

 

[RebateMonger]

The one other "easy" fix would be if you are a victim of Seagate's firmware flaws that would "lock up" the 7200.11 drives. Looking at your Firmware versions, I doubt this is the problem, but you should look at Seagate's support site and their forums to make sure you have the "good" firmware versions.

 

[zefyx]

Not a firmware issue as far as I know, I typed the serial into Seagate's verification page & was told no updates were available. Upon further investigation I found that the drive in question does make a 10 second noise upon power up that none of my other drives make. The weird thing is, I've never heard this noise with hard drives before. It's something along the lines of.. long... click click long... click click & then it disappears.

I've updated the original post with a screenshot of HDTune in regards to this drive.

 

[Emulex]

intel matrix raid is poop.

raid-5 without a real bbwc controller is poo.

using non ES or NS (RE4) drives for raid is poo.

you can mark the drive normal in the bios which will make it pretend to be good again and try to recover.

welcome to how much sata raid (with consumer non raid drive) sucks club!

 

[MerlinRML]

Originally posted by: zefyx
I tried then tried switching in my bios from RAID to IDE to run Seagate's SeaTools to test the drives, SeaTools would not run & gave me & error which wouldn't allow it to run. I then rebooted the PC & switched my drives back to RAID. At this point, Intel's RAID boot-up told me my array named 4.5TB had failed & that one of my disks was no longer listed as a member of the array

Some information:
Windows XP Professional 64-bit Edition
(3) 1.5TB Seagate HDDs
GPT formatted disks

So I'm quoting the interesting things that I see and it's definitely generating a couple of questions.

First of all, what are you booting from? I don't believe Windows XP x64 can boot from GPT disks. And while the boot disk in question may not make a difference to the failed RAID 0, it may make a difference how you got into the position you're in.

Secondly, when you say you booted in IDE mode to run seatools, did you boot into DOS and run seatools, or did you boot into Windows and try to run seatools? If you booted into Windows, there's a really good chance that it tried to read the partition information on each of the disks that make up the RAID array and got the disks out of sync. Thus, your RAID array was corrupted because Windows treated them individually.

 

[zefyx]

Thanks for the reply/

So I'm quoting the interesting things that I see and it's definitely generating a couple of questions.

First of all, what are you booting from? I don't believe Windows XP x64 can boot from GPT disks. And while the boot disk in question may not make a difference to the failed RAID 0, it may make a difference how you got into the position you're in.

Secondly, when you say you booted in IDE mode to run seatools, did you boot into DOS and run seatools, or did you boot into Windows and try to run seatools? If you booted into Windows, there's a really good chance that it tried to read the partition information on each of the disks that make up the RAID array and got the disks out of sync. Thus, your RAID array was corrupted because Windows treated them individually.

I'm booting from a Non-Member Disk which is formatted to NTFS. I booted into DOS & Windows when trying to run SeaTools.

 

[MerlinRML]

Originally posted by: zefyx
Thanks for the reply/

I'm booting from a Non-Member Disk which is formatted to NTFS. I booted into DOS & Windows when trying to run SeaTools.

Booting to DOS in IDE mode should have been fine. Booting into Windows in IDE mode will do bad things to your RAID array.

I'm not 100% sure what happens, but I've seen this effect before. Windows booted and loaded its IDE driver. It saw all of your RAID member disks as individual disks and tried to read each one. From there, I'm not sure how or why, but Windows updates some data on one or each of the disks. This is where your problem comes from.

When Windows modified the data on one of your disks, it was no longer in sync with the other RAID disks. So when you switch back to RAID mode next time, the RAID BIOS sees that the data has been touched and kicks the disk out of the array and fails/degraded the array. I think that most of your data should be ok, but the trick is finding the right commands (if there is one available) to force the disk back online and back into the array. You can deal with whatever corruption you have from there.

As to what your initial problems were that started you down this path in the first place, don't know if you'll be able to figure those out in the current state.

 

[glugglug]

Originally posted by: zefyx
Not a firmware issue as far as I know, I typed the serial into Seagate's verification page & was told no updates were available. Upon further investigation I found that the drive in question does make a 10 second noise upon power up that none of my other drives make. The weird thing is, I've never heard this noise with hard drives before. It's something along the lines of.. long... click click long... click click & then it disappears.

I've updated the original post with a screenshot of HDTune in regards to this drive.

I've had a few WD drives die this way and seen Quantum drives die this way before aka the click of death.
Basically the armature is stuck in the parked position and can't make its way out to read the drive. A UCF computer store tech told me Apple used to charge customers $45 to fix this when some models of Quantum drives had it as a really common problem using the "Apple Torsion Restart" which basically means they hit the side of the drive with a hammer to get it unstuck (with actual specs on what kind of hammer to use and how hard to hit - not recommended until you have exhausted other options).

If the drive just failed to unpark on one reboot causing the RAID array to remove it, and it isn't really stuck, you may be able to remove and re-create the array in the BIOS. Be sure to note what order the drives go in the array first.

 

[Syntax Error]

First thing I could say is that I hope you kept backups and there was no mission-critical data on that array.

If one drive in a RAID0 array fails, the entire array is toast, RAID0 actually INCREASES the chances of data loss due to its fragility through striping. The best thing I could hope for is that the drive that is failed (which appears to have timed out or has self-bricked, going by the infamous 7200.11 firmware fiasco) will be recoverable and be admitted back into the array with no data loss, but that's a bit of a long stretch.

I don't trust the Seagate verification tool, it's been known to be wrong before of drives that have had issues freezing in RAID arrays, timing out, or are bricked on bootup.

 

[VirtualLarry]

Originally posted by: zefyx
The worst part of all of this is I have another 1.5TB drive that is supposed to be delivered tomorrow, I ordered this drive so that I may convert my RAID0 array to RAID5 for redundancy.

That's too bad. You learned a hard lesson - you shouldn't store any data onto a RAID0 array that you cannot afford to lose at a moment's notice.

 

[zefyx]

Just an update. I ended up chalking up the data loss, in an attempt to not spend the next month crying over spilt milk. I've added another 1.5TB Seagate & converted to RAID5. The odd thing though, how did I loose it in the first place. My assumption is the drive timed out, because the same drive is doing fine in the current RAID5 setup. The only thing out of place is what sounds like the HDD's arm clicking, which I still can't explain. I know the reason why its doing it, technically but I'm not sure why it is having too. Thanks for the help.

 

[RebateMonger]

Sorry you lost your data.

You have more courage than I. You've got a drive that makes funny noises and has already been dropped from the array once. At the very least, I'd run a couple of day's worth of extended disk diagnostics programs from various makers, plus several full surface scans.

Actually, the truth is I wouldn't use the drive. I still have a brand-new 160 GB Seagate drive that dropped from a new RAID 1 array three times. I replaced it with another drive, which has functioned perfectly in RAID 1 for five years now.