RAID 5 failed on one drive

[bosshogg]

Hi Guys.
I have a RAID 5 made up from 3 500GB drives. Last week one drive died. The RAID was "degraded" but still accessable. I removed the faulty drive and then the RAID became "Failed" and inaccessable. "Failed" is a status for 2 or more dead drives but in my case only 1 drive is dead, even the RAID software reports this. I've received the new replacement drive and installed it but my RAID is still "Failed".

Because of the "Failed" status auto recovery did not kick in. The instructions in the Intel Matrix 7.8 tells that I must delete the exisiting volume and then created a new one.

I know all my data is still there on the 2 original drives but what do I have to do? Why is it reporting "failed" when the other 2 drives are still good?

 

[boonality]

The only thing i can think of is that maybe you pulled a working drive on accident and the bad one is still in the array?

 

[evongugg]

RAID Reconstructor
Recover Data from a broken RAID Level 5 or 0 Array

http://www.runtime.org/raid.htm

 

[bosshogg]

Thanks guys.

Some after thoughts...

When the drive died the RAID went into "degraded" mode. I removed the drive physically and rebooted this is when the RAID went into "Failed" mode. I put the faulty drive back in but it remained in "Failed" mode.

Maybe you're not supposed to touch it until you can perform the swap but in my case I had to return the drive first to get a replacement.

I removed the drives based on the serial numbers reported in the application and all appeared correct.

 

[boonality]

Ya could be, perhaps hot swapping would have prevented the whole situation, but how do you hot swap if you have to send a drive in for a replacement... leave your computer running for a week or 2? What if the power goes out? I hope you can recover everything

 

[chunkymonster]

Maybe you're not supposed to touch it until you can perform the swap but in my case I had to return the drive first to get a replacement.

Yes. Generally speaking you do not want to remove the faulty drive until you can swap it out with a good one. You don't have to and can replace the faulty drive with a good one at anytime then rebuild the array, but as a best practice it's best to remove and replace the drives at the same time.

If the RAID utility does not let you rebuild the array, then as already suggested, RAID Reconstructor from www.runtime.org is an excellent tool to recover your data with. I've used it with much success.

 

[bliq]

interesting- in linux land, you can mark a drive bad, remove it, and then mark it good, so it starts to recover. Maybe there was a step you missed where the faulty drive needed to be marked bad before you removed it.

 

[bosshogg]

I'm gonna work with RAID constructor now as best I can. The next hurtle might be that I have 600GB over 3 500GB drives. If RAID constructor produces an image where will I put it? Here goes...

 

[infernoxca]

i guess now we all know why people use hot spares

 

[bosshogg]

what I don't get is why it dropped from "Degraded" to "Failed" just because I removed the drive. It was still so after I put the drive back in.
Now I can still see that the two original drives are fine but they won't rebuild to the 3rd.

The RAID reconstructor tool is not helping either. Just waiting on Intel Matrix support to reply before I put a gun to it's head.

 

[rockchalk]

With Intel Matrix Storage, you actually could have tried to mark the disk as non-failed and tried to rebuild to it. You can right-click on the failed drive and select something like "mark as normal".

The only way the volume should have gone failed would be if you pulled the wrong disk. Port numbering is not always obvious on motherboards, and it can be easy to pull the wrong cable.

If you were able to get into the Windows UI (i.e. via another boot disk), and you had re-inserted the missing disk, I believe you could also right-click on the volume and select "Recover Volume" which would mark it as healthy again. I know this works with Intel's RAID0, and I believe it works with their other RAID levels as well.

 

[bosshogg]

Here's the robotic reply from Intel:
"Thank you for contacting Intel(R) Technical Support.

Once the RAID 5 volume is "failed", there is no way to restore it. The RAID 5 structure can handle only one drive failure. If more than one drive is set as failed or missing, the volume will collapse.
"
All I can think of is I must have dislodged a lead to one of the good drives on the next reboot. It's not hard to do as they don't click into place too good.

Rockchalk thanks for commenting. I don't have those options at all, believe me I looked!


I've decided that the drive data is written off but now I'm really stuck for how to rebuild them again. Get 4 drives and go for RAID0+1 or fcuk RAID altogether and just back up to a seperate drive and DVDs everyday?