Revodrive + Asus Sabertooth X58 + Intel Raid 10

[melacon]

After installing a new Revodrive, successfully into Windows 7 Pro, my Intel Matrix Storage console is reporting that two of my four Raid 10 drives are "missing hard drive". But the two drives are listed under the "Non-raid drives" section. I cannot figure out why they have been removed, or how to restore the volume.

I tried removing the revodrive and booting without it, but during POST, the Intel controller is reporting the same thing. There are no options to restore, reinitialize, etc. Reinstalling the revo drive changes nothing.

Thinking that all my life's critical data was safe on a Raid10, i have over 2 TB of data there that I really would like to get back. I haven't done anything yet to the volume that would add to it, hoping there is some way of rebuilding the volume somehow. The data is all there, but for what ever reason the Intel controller doesn't think the two drives are part of the raid.

I have been looking at raid recovery software to figure this out. Or maybe a recovery service. Any help out there?

Attachments

[einstein9]

i guess you should contact Dr.Kiev on RAID issues. he is the master of RAID btw here

good luck

[melacon]

Using Winhex, I can open the two drives that appear to have ben dropped from the volume. I believe they are the same half of each side of the volume, which is bad news.

Here is the worse part. The part of the raid that is still "intact" seems to have been corrupted. When I use Winhex (or any other program) to access it, the program freezes for a long time while trying to read from it. Eventually, it will start reporting back that there are I/O errors and corrupted sectors along the side I think a data recovery service is going to be my only bet here...

Attachments

[quasimodo]

send it to a DR lab, or at least create an image copy of all drives before attempting anything else.

[melacon]

$5000 to $7000 quote on the DR lab.

[melacon]

I had tried all sorts of combinations with the BIOS. IDE, AHCI, Raid, with the revo in, took it out, some drives in, some out. No change. My first run of Raid Reconstructor was "Not significant" or something, so I bumped up the sector scan size to 500,000. The log also gets really full really fast of messages "unable to read sector, entering recovery mode, successful". So, it seems that something has happened to mess with the integrity of the data on part of the volume. I'll see what it says after a scan at 500k, and go from there. My question is this, now. The raid controller still thinks there is a raid 10 hanging out somewhere. This is causing the computer, and RR, to see one of the physical drives as a 4tb drive, or something like that. I'll post a SS. The question is, would it help or hurt to "break" the raid at this point? I mean, the data will still be there, in some form, I think. Does it give me any advantages or disadvantages? I know I can't go back once I do that, so I am reluctant.

Attachments

[drc]

Bluntly, raid reconstructor sucks.

If you image your drives then you at least know that you have a copy of everything before you start messing with it. At this point you don't even know if some of the drives are not working correctly, or have bad sectors, or whatever (I'm guessing no, but the fact that you are trying to access them through the raid controller instead of individually is clearly causing issues)

You said the two drives that have dropped out appear to be mirrors of each other? Then you just need one drive from the other mirror set, and some software that will emulate the raid setup and let you pull off the data (or image, or whatever)

[melacon]

See, I am REALLY new to this hdd failure stuff. I guess I"ve been lucky over the years. Are you saying it would be ok to 'break' the raid, then image the disks, then work from the images? I suppose I would need at least 2 more 2tb hard drives, unless it would be ok to format the other half of the raid for that use. I just didn't want to do anything that is permanant until I was sure that I was either A)screwed or B)good to go. What do you recommend I use to image the drives? I can only imagine there are tons of programs out there, so I would have no idea what would be the most practical.

[quasimodo]

your minimum of investment.

Stop what you are doing, purchase 4 New 1 TB drives. Image each drive to be on the safe side, before doing anything further. (don't make any modification to the raid). If the drives are good use Norton Ghost or some other tool to copy each drive individually. If you receive lot's of bad sectors or other problems during the image copy stop immediatley!

[zebong]

take attention if the NAS do not format your disk ....

i had a case from Intel that the NAS had been formated, also a raid10.

so is better you to clone all the drives first.

[melacon]

Why do you need to stop imaging if there are bad sectors? Isn't that the point, to read the data off to put it into a safe place?

If I image a 2tb disk, does it take a full 2tb disk to hold that? Or does it only go where there was data? I have my four new 2tb on the way, so I need to image all four bad ones onto them, I think. But then I have no where to put the good data when it's used, except where the 2 of the 4 from before that I suspect are ok.

[quasimodo]

yes if the drives are 2TB each then you should image the entire 2TB to a new disk each. Do not write to the old drives. If you do not have any more storage, then buy another drive or two. The cost to invest in some additional drives should be worth it to you.

Why do you need to stop imaging if there are bad sectors? Isn't that the point, to read the data off to put it into a safe place?

As far as a drive with many bad sectors, I would recommend you to stop imaging that drive because depending on the drive's condition you could further stress the drive and thereby possible damage it. In addition Ghost and other image tools can't handle bad sectors well. Yes, the point is to create an image of each drive and to preserve it, however the goal should also be to not cause any further damage to the original drives during the process.

[melacon]

So if I am getting bad sectors, I shouldn't be doing anything to the drives, and therefore am stuck into data recovery service?

[quasimodo]

RAID arrays are pretty sensitive to bad sectors or bad stripes. It really depends. You could potentialy try one of the HD vendors freely available tools to test for any SMART errors, this could give you a better idea regarding the state of the drives. Bad sectors can be of logical nature or of a physical one, so it will be somewhat difficult for you to judge if you don't have much experince in this field. Of course you want to use a coherent methodology, and image the drives without any bad sectors first. If in doubt or difficulties are being encountered it would be wise to send it to a DR Lab.

[melacon]

So, I think I'm on the right track...

Installed my old PS3 60gb sata drive that I had laying around after removing the revo. I reverted the Intel controller to IDE mode, and did a fresh install of Win7. After hours of install and updates (stupid hdd is sooooo slow), I installed the trial of RaidLabs Raid Recovery. Put in the settings and within a few minutes it was showing me the recoverable files that were available. It seems like most, if not all, of them should be viable. Looks like it will be the best $$$ I will ever spend on software...

I have to say thanks to you all for your assistance, and to hhansard on the ocz forums for pointing me in the right direction!!!

Attachments

YAY!