Rebuilding RAID 5

Hi. I am new to the forum and to RAID. Our Adaptec RAID 5 (2100S) has failed and we are running on 2 of the 3 hard drives. The hot spare doesn't seem to be working either.

The 2 working drives are on SCSI 3 and 2. The drive on SCSI 1 is shown as good in the RAID administration utility but is no longer part of the RAID. The drive on SCSI 0 has failed.

Does "failed" mean that the drive is broken or just that it is not part of the RAID?
Can I rebuild onto the drive that is still shown as good?
If something goes wrong during the rebuild will that destroy the data on the 2 remaining RAID drives?

The trouble is that having swapped the drives around so much trying to get the thing to work, I'm not quite sure which SCSI # was the position of the drive that failed originally (d'oh).

If anyone has exprience with the Adaptec 2100S that would be great.
Thanks for any help.
Tony.

This sounds like a controller error to me. I would try flashing the firmware on it. Check to make sure the drive that is reading "failed" is spinning. Is it Id'ing? If the drives spins up and it's assigned a SCSI ID, the drive itself should be fine. But, you mentioned swtching the drives around. I'll address that later.

Normally, you have to have three working drives in the RAID 5 for you to be able to rebuild. You mentioned only two are actaully functioning as part of the RAID. However, if it is giving you an option to rebuild, do it. There shouldn't be a problem with data loss if you just select the automated rebuild option.

I first thought drive 0 just plain failed and that would kick in the online spare. All you would've had to do was replace the bad drive and the array should've auto rebuilt. There might have been some prompts, yes, but it's pretty foolproof. Having switched the drives around is a problem indeed. I think the easiest solution might be to test each drive independantly first. Does your mobo have an onboard SCSI port? If it does use that. Hook up each drive and go through the POST. If one doesn't spin up, or if it won't ID, it's bad. Create a new RAID 5 with the three working disks you have. Get a new disk as soon as you can and just expand the array to it. Good Luck!

Thank you Mr E. That is extremely helpful. A couple of other questions:

Am I correct in thinking that for the RAID to rebuild, the drive onto which it is going to be rebuilt has to be in the same SCSI position as the drive that failed?

If the controller gives me the option of rebuilding can I take it that it "knows" the rebuild is safe?

When the controller shows a drive/SCSI position as "redirected" (I assume the spare) does that indicate that the drive is actually marked and operating that way or merely that the controller is programmed to have a drive operating that way at that position?

Tony.

Okay. The answer to the first quesiton is yes. I still don't see why it (rebuild) wouldn't work providing that you still have at least three drives online. The controller has likely seen the fact that you've switched the drive positions and is still using (or trying to use) the online space. The RAID is extremely fragile at this point, be careful.

Secondly, yes. The RAID controller sees that there is enough room to rebuild. It should be able to do it without data loss just fine. Again, I would try and rebuild as soon as you can. The RAID is teetering on the edge.

Lastly, I think the controller is seeing that the drives have been swapped and is waiting to rebuild to the online spare. It's at a different position than it thinks it should be, which is correct. It wants to use the third online disk as the hot spare, but its at the wrong position. You still have two of the drives' positions sequential, but it's confused about the third. That may be why you see the "redirected" message.

If you can still boot to Windows do so and make backups. RAIDs are tricky sometimes and all hell can break loose by making one wrong keystroke in the config menu. I hope I've helped!

Thank you so much for your help. The RAID is back to an optimal status and I've reconfigured the redirected drive to be the hot spare. End of a nightmare. Tony.

Hello,
I have a fair amount of RAID experience, but I am not sure what I am missing.
I have an Adaptec 3200S, i have a 5 drive RAID 5 array. The drive at ID 1,2 showed failed, both in bus and under array. Array showed degraded. I replaced the failed drive. Under normal operation I had drives 1,1-1,5. Now that I have replaced the faulty drive, it shows up as 1,2 as the original had, but now there is a 1,0 status missing. I never had a 1,0 until I replaced the failed drive. Also, now that I have an addition drive (ghost image of some sort??) that isn't actually there, the RAID section shows 1,1 optimal 1,0 missing, and 1,3-1,5 optimal, array status degraded, and no 1,2 which is what failed and was replaced to begin with.
To me this seems asinine, that when I replaced a drive and used identical SCSI ID's, that is generates a new ID for the drive that isn't there.

I have gone into the rebuild array section, and it does nothing at all when I select it.

I know I have fixed a problem like this some time ago, but, for the life of me I can't remember how I did it.

Thanks in advance!