Intel RAID5 Problem? Is HDD Dead?

[Zack.]

Hey,

I have a 6x1TB Raid5 Array on my Gigabyte UD5.

Just recently randomly during the day the intel raid manager popped up saying the RAID is degraded and one of the HDD has failed.

Now I checked and this is what it looks like:

http://img206.imageshack.us/img206/2858/64567843.jpg

It still has all the details of the drive and the drive still shows up?

I can right click and click "mark as normal" but I'm not sure if I want to do that if the drive is actually broken. Then that means if a different one dies and the one I marked "normal' was dead, I would have 2 dead drives and lose all my data.

What should I do? How do I know if that drive is actually broken?

Thanks.

PS. Removing it from the case is a lot of work so I would only want to remove it if its broken.

[enteon]

check SMART, nuke it and reinitialize?

... sounds like hollywood :ROTF:

maybe you want to think about raid 6? 2 failing drives allowed and you seem to have enough disks ^^

[Zack.]

Nooo I can't RAID6 because I have 5TB of data and I have no where to put 1TB of data lol.

WHat you mean SMART?

It's on the Intel RAID Controller.

[SoulsCollective]

Unplug all the other drives, plug in just the problem drive, boot into a Linux live distro and check the SMART data for the drive. That's like the drives internal record of which sectors are failing/going bad, but RAID controllers generally can't read data for individual drives and pass that through to apps that will read that SMART data for you so you need to be in single-drive mode.

Another option would simply be to swap the drive to a different port. If it's got problems, it'll be read as a failed drive on another SATA port as well.

[stevecs]

What SoulsCollective says is good, to check the smart data (smartctl) i.e. smartctl -a /dev/<device> to get a complete list. However I have found over the years that it's generally not too useful to keep on coaxing a drive back. It's better to just replace it or warranty it out if that's an option.

[Zack.]

Thanks guys.

I have done what you said and it seems like the drive was fine...

I just started the rebuild process and its at 25%.

Will let you know if it succeeds.

Thanks.

[Xcel]

I'd replace the HD if it drops out again and you value your data. I had that happen to me a HD dropped out maybe 5 times (SMART didn't say anything when I checked as a single drive) then all of a sudden the entire raid volume wasn't being recognised by the controller anumore, no matter what I did. I had to recover all the data manually using recovery programs and copying the data to other HDs. All that because of one dying HD...

[zanzabar]

i had a few problems with drives dropping at boot when i had the dfi x38 and the blackops from one of intels raid bioses, i would update the bios then see if it dose it again

[Speederlander]

This is why RAID 5 kind of sucks. The most likely failure point for a drive is on rebuild (especially as the array gets larger), and when rebuilding you can't afford to lose a drive without losing everything. I would never go RAID 5, at least not for an array larger than 1TB.

[Zack.]

I was sweating because I was so nervous when it was rebuilding! haha!

It's rebuilt 100% and everything is running fine. No loss of data!

Thank goodness. If it happens again I'll definitely replace the drive. I know which port its on.

Thanks for the help!