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 5-disk RAID-5 array in an SBS 2008 server, and it's been working fine for > 12 months now, but today I logged in to find this in the disk management console:

alt text

Of the five-disk array, it seems to have doubled itself up.

  • Array 1: Disk 1 & 5 are fine, Disk 2,3,4 are "Missing"
  • Array 2: Disk 2,3,4 are Foreign, disks 1 & 5 are "Missing"

Disk 0 and 6 are seperate, non-RAID disks

I don't have physical access to the server, but I have some remote hands and they tell me that all the lights on the disk bays are on, so I very much doubt that all three disks have failed.

I'm too scared to import the foreign disks, because 2 of the 5 disks are missing from the set, I don't want it to destroy the data on the array (Yes, I have backups, but that's a huge hassle).

Any ideas what might have happened here, and how I can can "glue" the array back together? I've already removed each disk from the device manager and then scanned for new hardware in case that fixed it, but it didn't, and I can't think of where to go from here.

share|improve this question

2 Answers 2

up vote 4 down vote accepted

According to this MS Article, grab your lucky rabbit's foot, cross your fingers, and Reactivate the Missing Disks. Then run chkdsk after they resync.

I've had this happen to me before on a single disk, but never 3 at once.

share|improve this answer
1  
Thanks. I gave this a go as it seemed to be the only option and they came back online phew. Running chkdisk and a prompt backup right now. –  Mark Henderson Oct 21 '10 at 3:21
1  
@Farseeker Hooray! Always nice when something like that works out. –  sysadmin1138 Oct 21 '10 at 3:54

Looks like you've got a mess.

If you haven't reviewed the System Event Log for events leading up to the failure I'd recommend you start there

It looks like you took failures on disks 2, 3, and 4, knocking them out of the array and altering their signatures such that they're being recognized as Foreign disks now. In old SCSI days I'd be wondering about an ID conflict between the host adapter and a target on the bus that these disks were connected to, but that doesn't explain why their signatures are screwed-up. (I'm assuming they're SAS, anyway, so that should be a non-issue.)

Importing the Foreign disks isn't likely going to help. Their signatures have been damaged and, to my knowledge, they're not going to just "jump back in" to the RAID-5 volume. I don't think you're going to have any luck "Reactivating" them, since their signatures have changed.

You're probably going to be breaking out the backup. If you do, run the manufacturer's diagnostic on the disks that experienced the "failure" before you put them back in production. (It seems a lot more likely that a controller bug would've caused 3 disks to hiccup simultaneously, but check the disks over anyway.) What kind of controller are these disks sitting on, anyway?

A brief editorial comment: Those are awfully big disks to be members of a RAID-5 array. Presumably those are SATA disks. The bit-error rate of the media could, potentially, catch up with you during a resync of an array that large. I'm sure it's been configured this way because the Customer needs lots of space, but they're walking on the edge of a sword with a RAID-5 composed of that many large disks.

share|improve this answer
    
Thanks for the ever-useful advice as usual. I made sure the backup tapes were mounted and hit the import button and hey presto, after about 30 seconds they came back online and started resyncing. We had a disk failure a few months ago and it took 24 hours to finish re-syncing so we'll see how it goes this time. Correct, they're SATA disks, and they're sitting on the on-board SATA controller of an Intel PS-series motherboard. –  Mark Henderson Oct 21 '10 at 3:21
    
After this episode (latest in a long series with this server) I'm going to suggest completely re-thinking their disk infrastructure anyway and maybe moving to an entry-level iSCSI device. –  Mark Henderson Oct 21 '10 at 3:24

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.