Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I have a Windows 2008 Server with a two-drive software RAID 1 (mirroring). Last night one of the drives apparently failed (just disappeared from the system) and I had to boot from the secondary plex in order to get it running again. I don't know which of the drives actually failed as upon shutdown/restart both were online.

Now the Disk Management says the volume status is "Failed Redundancy". If I try to reactivate it, it starts resynching but apparently fails somehwere in the process as it encounters a bad block on one of the drivers (MIGHT be the other one).

I'm very confused now; I'm not sure at all what is the current state of the array - does Failed Redundancy mean it is still performing mirroring (there are no yellow triangles next to disks unless I try to reactivate them)?

And most importantly, I have absolutely no idea if the new data (the one that was created between last night and now) is contained on both drives or just one (which one?).

Any suggestions would be appreciated.

Thanks! Darko

UPDATE

I have figured out which drive contained bad blocks, but unfortunately (of course, when things go wrong, they go really wrong) it was apparently the other drive that got dropped out of the mirror, so I had the latest data on the drive with bad blocks and no way to resynch the mirror as Windows really doesn't want to replicate data from a drive with bad blocks.

So after I made enough backups I decided to give it one more try (before cloning the drive offline) and tried an application called "HDD Regenerator", which is supposed to somehow "regenerate sectors which are not physically damaged but just incorrectly magnetized" (note that I don't fall for shiny interfaces of applications that are supposed to do miracles) and to my amazement in my case it actually somehow worked - no sectors were relocated and the data is still there. After running chkdsk with /B parameter to recheck (52 KB previously) the bad clusters it reported none were found, so I was able to successfully resynchronize the mirror and verified that all the data is intact.

Of course I will be very closely monitoring the drive (it passes all diagnostic tests now so I doubt I can return it) in the future and will toss it in the garbage at the next sign of abnormality.

Thanks for all the help.

UPDATE a year later

Still working perfectly with no bad sectors.

share|improve this question
    
Not that I'm dubious or anything.... :) but when you say you'll be very closely monitoring the drive - monitor it very very closely! This smells like the software you ran just reset the SMART values on the disk. –  Ben Pilbrow Dec 5 '10 at 23:34

3 Answers 3

up vote 3 down vote accepted

"Failed Redundancy" means that one of the disks in the mirror has failed.

I think Bart's answer is a little bit extreme. Here's what I'd do:

  • Take a full backup of the server now.

  • Pull the "primary" disk out of the server computer, label it as to its connection to the server (which port it's connected to, etc).

  • Boot the server on the remaining disk.

  • Using "Event Viewer", verify that this is the disk you've been running on since the failure (i.e. look for entries that correspond to the time you've been up and running).

  • If the secondary disk is the one you've been using replace the primary disk then proceed.

  • If the secondary disk does not appear to be the one you've been using pull it out of the server, install the primary disk, and repeat this process. You need to be sure that you're working with the disk that you've been running on.

  • Test both disks with the disk manufacturer's diagnostic tool and determine which disk has a problem. If the disk you've been using is the problem disk you probably won't be able to re-establish the mirror (because Windows will stop mirroring when it hits a bad block on the source disk) and you'll be best off restoring your backup to a fresh disk. If the disk that failed turns out to be the disk that you haven't been using since the failure occurred then you'll probably be able to get a fresh disk and re-establish the mirror.

I'd be very wary of booting the server with both disks installed until you've established a clear understanding of which disk failed and whether or not you'll be able to re-establish the mirror.

share|improve this answer
    
Hey! I like my answer. :-p –  Bart Silverstrim Dec 3 '10 at 17:46
    
@Bart: I don't dislike it... >smile< I just think it's worth trying to rebuild the mirror if all that's needed is a fresh disk versus immediately flattening the box and starting over. –  Evan Anderson Dec 3 '10 at 17:56
    
@Evan...yeah...well...I still suck my thumb in your general direction :-) –  Bart Silverstrim Dec 3 '10 at 18:59
    
Thanks for the answer. I have updated my original post with the "aftermath". –  Carko Dec 5 '10 at 23:09

Data on the drive that failed is not reliable, and shouldn't be counted in for redundancy. Assume that drive has no good data on it. Per Microsoft documentation:

A mirrored volume's status is Failed Redundancy.

Cause: One or both of the members of the mirrored volume has failed, and the volume is no longer fault tolerant. To avoid data loss, you should attempt to repair the volume as soon as possible.

Isolating which drive is the bad one in the event that you lack appropriate blinknlights is doable. Unplug one and try to boot from it. If it works and you have data, you found the good one. If it doesn't work, you have found the bad one. Unplug it, plug the other one back in, and go find a new drive.

share|improve this answer

Depending on the value of the data, I'd make a backup from the server now, wipe the drives with a fresh format to see if one of them errors out, and recreate the mirrored volume and restore the data.

Usually the failed redundancy is just that. You have one drive that isn't synced up, and you're running with one primary drive working properly.

I'd wonder if doing the backup-reformat will help you discover which individual drive has a bad spot or oddball cluster and then replace it.

share|improve this answer
    
Thanks! I have figured out which drive contains bad blocks (did the surface scan on both). I'll first try to boot just from the other one and if that won't work, will try to somehow create & restore the backup of the entire volume... –  Carko Dec 3 '10 at 17:40

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.