On Thu, Sep 24, 2009 at 5:03 PM, MarvinC <marv...@gmail.com> wrote: > 2. There's a Simple 36gb drive.
That could be a hot spare. > Sooooooo! For some reason the server errors and show the drives as Failed > when attempting to boot from the SCSI controller card. That sounds like it has a hardware RAID controller card, not a plain SCSI card. With a plain SCSI card, a failed drive normally either doesn't appear in the SCSI BIOS POST list, or hangs the SCSI BIOS detection probe, or fails to boot. What's the *exact* message it's giving you? > If I change the cable I get a different error. What's the "different error"? > If I take the drives and add them into the other, > identical, server, I get NVRAM mismatch errors on Post ... That also sounds like a hardware RAID controller. The controller is saying the configuration in NVRAM on the card does not match configuration on the disks. > So I take the cable and plug it into the onboard SCSI connector. Eeek! That's bad. As in, "don't cross the streams" bad. The OS will likely do things to the drives without following whatever protocol the RAID controller uses, effectively ruining the RAID set. > The servers boots into the OS, sees the drives but lists them as Failed > and Foriegn. Right, because a mirror managed by a hardware RAID controller will generally appear like two identical copies of the same disk. The second will appear as foreign because it has a duplicate signature and the OS won't tolerate that. The RAID 5 set will be completely unrecognizable to the OS. > When I attempt to Reactivate Volume or Import either of the > Foreign disks I get the following error in the System Log: RAID implemented by a hardware controller and RAID implemented by the Windows OS are two completely different things. There's as different as FAT and NTFS, or NTFS and Macintosh HFS+. The OS has no idea what to do with your hardware RAID sets. > My gut tells me the system's hosed and that all data on the dynamic drives > is lost ... A hardware RAID 5 set isn't a "dynamic drive" or anything else from Microsoft. Without the proper RAID controller, it's an opaque entity which Windows cannot use. You might have been okay before you started messing around with the RAID sets like that. Doing that quite possibly scrambled the contents of the RAID 5 set. It certainly broke the mirror set. I know this is harsh, but since this is a data loss situation: You should contact someone who knows what they're doing. You have almost certainly made things much worse by flailing around like this. I'm sorry if telling you this hurts your feelings, but if you're a good tech you'll recognize that the client's data is more important than your pride. There are data recovery companies which specialize in fixing this kind of disaster. OnTrack is probabbly the most well-known. I've also used CBL <http://www.cbltech.com/> with success. Tell them what you did. They'll probably want you to ship all the disks *and* the RAID controller to them. It will be expensive. If you insist on doing it yourself, or if the client can't afford to get themselves out of this hole: Since you've got a sort-of running system, first copy off everything from the one mirror member that you can, while you can. Next, I would recommend creating block-level images of all the existing disks, using another computer, and a plain SCSI controller. That way you have a backup if you make things even worse than they already are. I would do this by booting Linux from CD, attaching a disk with at least 200 GB of free space, and then using "dd" to image each disk. For example: mount /dev/sda1 /backup dd if=/dev/sdb of=/backup/disk1.img dd if=/dev/sdc of=/backup/disk2.img dd if=/dev/sdd of=/backup/disk3.img dd if=/dev/sde of=/backup/disk4.img dd if=/dev/sdf of=/backup/disk5.img (The above assume the first disk, "sda", is the data disk with the filesystem you're storing the backups on.) Then I would shut that down, and put all the disks back in the original computer, with the original RAID controller, *EXCEPT* I would only put in one member of the mirror set. Since you're booted and run off one of those, the other member will be out of sync. If you boot with both, I have no idea what will happen, but "scrambled filesystem" seems likely. Then try to convince the hardware RAID controller to recognize the old RAID-5 set again. If you're really lucky, the configuration-on-disk hasn't been blown away, and you'll be able to get it to mount again. Possibly degraded with one member failed. Then copy everything you can off the system. If you can't get the controller to recognize the RAID 5 set from the disks, see if there's an option to re-create the RAID set in degraded mode without reinitializing all the disks. If you're really lucky, you can find a way to tell the controller the way things were, and then the data will still be there. See if the company who made the RAID controller is still around. Their tech support may be able to help. -- Ben ~ Finally, powerful endpoint security that ISN'T a resource hog! ~ ~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/> ~
